1 Imaging suicidal thoughts and behaviors: a comprehensive review of two decades of neuroimaging studies Lianne Schmaal*1,2, Anne-Laura van Harmelen*3, Vasiliki Chatzi3, Elizabeth T.C. Lippard4, Yara J. Toenders1,2, Lynnette A. Averill5, Carolyn M. Mazure6, Hilary P. Blumberg7 * shared first authorship 1Orygen, The National Centre of Excellence in Youth Mental Health, Parkville, Australia 2Centre for Youth Mental Health, The University of Melbourne, Melbourne, Australia 3 Psychiatry, University of Cambridge, Cambridge, UK 4 Psychiatry, Dell Medical School, University of Texas, Austin, USA 5 Psychiatry, Yale School of Medicine, New Haven, USA and Department of Veterans Affairs National Center for PTSD, Clinical Neurosciences Division, West Haven, USA 6 Psychiatry and Women’s Health Research at Yale, Yale School of Medicine, New Haven, USA 7 Psychiatry, Radiology and Biomedical Imaging, and the Child Study Center, Yale School of Medicine, New Haven, USA Corresponding author: Name: Hilary Blumberg, M.D. Affiliation: Yale School of Medicine Address: Department of Psychiatry, 60 Temple Street, Suite 6B, New Haven, CT 06510 Telephone: 203-785-6180 Email: hilary.blumberg@yale.edu Word count: Text: 9386 words, 3 Figures, 3 Tables 2 ABSTRACT Identifying brain alterations that contribute to suicidal thoughts and behaviors (STBs) are important to developing more targeted and effective strategies to prevent suicide. In the last decade, and especially in the last 5 years, there has been exponential growth in the number of neuroimaging studies reporting structural and functional brain circuitry correlates of STBs. Within this narrative review, we conducted a comprehensive review of neuroimaging studies of STBs published to date and summarize the progress achieved on elucidating neurobiological substrates of STBs, with a focus on converging findings across studies. We review neuroimaging evidence across differing mental disorders for structural, functional and molecular alterations in association with STBs, that converges particularly in regions of brain systems that subserve emotion and impulse regulation including the ventral prefrontal cortex (VPFC) and dorsal PFC (DPFC), insula and their mesial temporal, striatal and posterior connection sites, as well as in the connections between these brain areas. The reviewed literature suggests that impairments in medial and lateral VPFC regions and their connections may be important in the excessive negative and blunted positive internal states that can stimulate suicidal ideation, and that impairments in a DPFC and inferior frontal gyrus (IFG) system may be important in suicide attempt behaviors. A combination of VPFC and DPFC system disturbances may lead to very high risk circumstances in which suicidal ideation is converted to lethal actions via decreased top-down inhibition of behavior and/or maladaptive, inflexible decision-making and planning. The dorsal anterior cingulate cortex and insula may play important roles in switching between these VPFC and DPFC systems, which may contribute to the transition from suicide thoughts to behaviors. Future neuroimaging research of larger sample sizes, including global efforts, longitudinal designs, careful consideration of developmental stages, and sex and gender, will facilitate more effectively targeted preventions and interventions to reduce loss of life to suicide. Keywords Attempted Suicide; Suicidal Ideation; Suicide; Neuroimaging; Functional Neuroimaging, Magnetic Resonance Imaging 3 1. INTRODUCTION 1 Around one million people die by suicide annually1. Globally, suicide is the tenth leading 2 cause of death for all ages and the second leading cause of death among young people aged 3 15-29 years1. In 2013, it was estimated that 9.3 million adults 18 years and older in the United 4 States had suicidal thoughts and 1.3 million attempted suicide2. Additionally, a 2011 report 5 estimated that 13% of adolescents planned a suicide attempt in the previous year and 8% 6 attempted suicide3. Unfortunately, suicide death rates have continued to rise. For example, 7 since 1999, rates in the U.S. have increased 30%4. Predictive value of currently identified 8 non-biological risk factors for suicide is limited5, and a reliable biological risk marker has yet to 9 be identified. In order to prevent suicide more effectively, there is an urgent need to better 10 understand the mechanisms that confer increased risk for suicidal thoughts and behaviors 11 (STBs), and to identify biological markers of risk to generate more targeted successful 12 prevention strategies and monitor responses to them. In the last decade, and especially in the 13 last 5 years, there has been exponential growth in the number of neuroimaging studies 14 reporting structural and functional brain circuitry correlates of STBs (Figure 1). In the last 10 15 years, a number of excellent reviews on aspects of this research have emerged6–10. Here we 16 review research across structural, functional and neurochemical neuroimaging modalities, 17 providing a narrative review of 130 neuroimaging studies with a focus on the most researched 18 brain circuitries and findings that converge across studies. 19 20 2. METHODS 21 A search was performed in PubMed for original research articles published before March 12, 22 2018. The following terms were used: “MRI”, “SPECT”, “PET”, "magnetic resonance 23 imaging", "positron emission tomography", "single photon emission computed tomography", 24 “DTI”, "diffusion tensor imaging", “diffusion weighted imaging”, “neuroimaging”, “fMRI”, 25 “functional magnetic resonance imaging”, “spectroscopy” (separated by OR) in combination 26 with the terms “suicide”, “suicidal”, “suicidality” (separated by OR). We selected articles that: 27 (i) were published in a peer-reviewed journal in English and (ii) included groups with suicidal 28 ideation (SI) and/or history of suicide attempt(s) (SA). Of note, non-suicidal self-injury (NSSI) 29 was not included in this review, and studies that did not differentiate NSSI from suicidal 30 behaviors were excluded, because these can be differentiated on the basis of intention, 31 frequency, and lethality and may have partly distinct underlying mechanisms11. 32 33 3. RESULTS 34 We identified 131 unique articles meeting review criteria (Tables 1-3). The populations 35 studied reflect reports that the majority of people with STBs have a diagnosable mental 36 illness. Major depressive disorder (MDD) or bipolar disorder (BD) account for over half of 37 suicide deaths12. After mood disorders and borderline personality disorder (BPD), the 38 prevalence of suicide deaths is highest among people with substance use disorders and 39 schizophrenia (SZ), followed by post-traumatic stress disorder (PTSD) and anxiety 40 4 disorders12.The mental disorders researched varied by study, with each study typically 41 including a single disorder, and the majority of studies conducted in individuals with mood 42 disorders. Most studies compared people with a mental disorder and a history of suicide 43 attempts (suicide attempters, SAs) to people with a mental disorder and/or healthy controls 44 (HCs) without a history of attempt. Fewer studies focused on SI. Most studies employed a 45 cross-sectional design and a single structural or functional imaging modality. The majority 46 were conducted with adults; only a small proportion with adolescents (Figure 1). A subset of 47 studies provided preliminary findings of associations between neuroimaging measures and 48 key risk factors for suicide, e.g. medical lethality of prior attempts, emotion dysregulation, 49 anhedonia, impulsiveness, and reduced cognitive control (for reviews see13–15). 50 51 Despite modest sample sizes of studies, and the heterogeneity of their clinical samples and 52 neuroimaging acquisition and analysis methods, converging evidence is emerging to support 53 roles for specific brain regions/circuitries in STBs. These are particularly in cortico-54 striatolimbic systems that subserve emotion and impulse regulation and include prefrontal, 55 cingulate and insula cortices, amygdala, hippocampus, thalamus and striatum regions 56 (Tables 1-3). Within the prefrontal cortex (PFC), studies vary widely in the selection of regions 57 studied, with regions of interest (ROIs) often including overlapping regions that encompass 58 ventral and dorsal, medial as well as lateral, PFC. Thus, while we identify the importance for 59 future study of specific PFC subregions, given their differing connectivity, cellular and 60 molecular features and functions, we discuss the PFC grouped broadly into the ventral 61 prefrontal cortex (VPFC; divided into medial and lateral portions), which has the highest 62 concentration of reported findings, the dorsal prefrontal cortex (DPFC; divided into lateral and 63 medial portions), and the anterior cingulate cortex (ACC). We also discuss findings in the 64 insula, and mesial temporal (hippocampus, amygdala), subcortical (basal ganglia, thalamus) 65 and posterior regions (posterior cingulate cortex, lateral temporal lobes and cerebellum). See 66 Figure 2 for definitions of the brain regions. Since the VPFC, DPFC, ACC, insula, mesial 67 temporal, basal ganglia, thalamus and posterior regions have been studied most frequently in 68 relation to STBs and because most converging evidence exist for the involvement of these 69 regions, we specifically focus our review on findings within these brain areas, as well as in the 70 connections between them. However, additional studies and positive and negative findings 71 not discussed below can be found in Tables 1-3. 72 73 We focus our discussion below on findings that are primarily derived from comparisons of 74 individuals with a mental disorder and STBs versus individuals with a mental disorder without 75 STBs (diagnostic controls, DCs), rather than comparisons with HCs, unless otherwise 76 specified. Findings based on a comparison between DCs and individuals with the same 77 mental disorder plus STBs are more commonly reported in the literature and are more likely 78 to reflect specific effects of STBs in that disorder, whereas comparisons with HCs may 79 include more general effects of having a mental disorder. Below we first detail structural, 80 5 functional and neurochemical findings within regions and end each section on a region with a 81 summary. We then follow with a section devoted to studies of connectivity among regions 82 within major implicated brain systems. 83 84 Ventral Prefrontal Cortex (VPFC) 85 The lateral VPFC (VLPFC) refers to the inferior lateral areas of the frontal cortex 86 encompassing lateral orbitofrontal cortex (OFC, brodmann area (BA)47, lateral BA11), inferior 87 frontal gyrus (IFG, BAs 44, 45), and lateral aspects of the rostral PFC (RLPFC, lateral BA10) 88 (Figure 2). The lateral VPFC plays a key role in cognitive control, including response 89 inhibition, and is activated when behavioral responses are modulated in response to the 90 emotional or motivational context16,17. The medial VPFC (VMPFC) refers to the medial OFC 91 (medial BA11) and medial aspects of the rostral PFC (RMPFC, medial BA10) (Figure 2). The 92 medial VPFC has a well-established role in self-reflection18, appraisal of internally-generated 93 emotions (both positive and negative)16,19, appraisal of past and imagined future events and 94 reward processing20,21. Structural, functional and neurochemical alterations in these regions 95 have been associated with maladaptive strategies for regulation of negative affect (e.g., 96 rumination), negative selfreferential thinking22,23, and diminished positive affect (e.g., 97 anhedonia)24. Evidence is mounting that emotion dysregulation has a central role in the 98 generation of STBs. This includes elevations in negative, and blunting in positive, subjective 99 emotions, self-referential thoughts, and responses to valenced stimuli25. These alterations are 100 thought to contribute to key clinical risk symptoms for STBs including depression, anxiety, 101 rumination, guilt, reduced self-esteem, helplessness, anhedonia and hopelessness22–24,26–31. 102 103 VLPFC 104 Structural MRI studies have consistently shown lower gray matter volumes of the VLPFC in 105 adult SAs, including SAs with MDD32–34, BD35,36 and schizoaffective disorder (SZA)37. 106 Findings of lateral OFC volume decreases, extending to medial OFC, in adolescent and 107 young adult SAs with BD36, suggest these may be early differences. Lower VLPFC thickness, 108 but not volume, is one of the rare findings related to SI in adults with MDD38, suggesting the 109 VLPFC may be involved not only in suicide behavior but also in the ideation that may 110 generate it. 111 112 Gray matter volume decreases in VLPFC were also associated with high lethality of prior 113 suicide attempts in MDD24 and BPD36,37. This converges with longstanding findings from both 114 postmortem studies demonstrating VPFC (ROIs including both VLPFC and VMPFC) 115 differences in people who died by suicide39, and positron emission tomography (PET) studies 116 of high lethality or high intent SAs, showing VPFC (VLPFC and VMPFC) alterations in 117 serotonin (5-HT) synthesis, transporters and 5-HT1a receptors40,41 (Table 2). Lower 5-HT1a 118 binding in the OFC was associated with interim SI during a 2-year follow-up in adults with 119 MDD and past attempts42. Although conflicting findings exist43–46, these data provide some 120 6 consistent findings of a location and a potential mechanism, i.e. lateral extending to medial 121 VPFC serotonergic dysfunction, as a potential biomarker of risk for high lethality STBs. There 122 have been few molecular imaging studies of other neurotransmitter systems and 123 neurochemicals implicated in STBs. 124 125 Functional MRI (fMRI) studies performed while subjects conducted specific behavioral tasks 126 provide evidence that STBs are associated with VLPFC functional abnormalities in response 127 to emotional and other hedonically-valenced stimuli. Increased activation of the IFG and 128 lateral and medial OFC while viewing angry (but not happy, sad or neutral) faces was 129 reported in adult SAs with MDD47,48. Higher IFG activation in response to angry faces was 130 also associated with poorer attempt planning and higher impulsivity in adult SAs with MDD49. 131 Furthermore, young adult SAs with BPD displayed higher lateral OFC activation while 132 instructed to experience and regulate negative autobiographical memories50. Increased 133 activation in a region of interest that included both the medial and lateral OFC was also seen 134 in response to winning a reward47 in adult SAs with MDD. 135 136 The IFG also plays a critical in cognitive control and response inhibition51. During 137 performance of a continuous performance task, higher IFG, RLPFC and lateral OFC 138 responses were associated with both attempts and SI in adults with mood disorders with 139 psychotic features, in the absence of task performance differences52. Higher lateral OFC was 140 also reported during error trials in a response inhibition task in veterans with SI53. In contrast, 141 a second study in adult SAs with SZ using the same continuous performance task showed 142 reduced activation in a cluster encompassing the RLPFC and IFG, extending to the VMPFC 143 and ventral ACC, that was associated with SI but did not further distinguish between ideators 144 with and without a history of attempt54. 145 146 VMPFC 147 In addition to the structural and PET study findings reported for the VLPFC above that 148 extended to the VMPFC or that were based on an ROI including both VLPFC and VPFC, 149 lower medial VPFC cortical thickness was also associated with greater motor impulsivity in 150 adolescent SAs with MDD55. As cortical thickness and surface area contributions to volume 151 are thought to be genetically independent56 and result from different neurobiological 152 processes57, it is important to examine these separately in studies of STBs. However, the 153 majority of studies have either not examined thickness and surface area separately from 154 volume or only thickness but without examination of surface area in SAs with MDD32,58. 155 Cortical thickness is thought to be influenced by the number and the size of cells within a 156 column, packing density, as well as by the number of connections and the extent of their 157 myelination, while cortical surface area is driven by the number ontogenetic columns that run 158 perpendicular to the surface of brain59. 159 160 7 Functionally, in addition to higher activation in the lateral and medial OFC in response to 161 angry faces and to winning a reward in adult SAs with MDD47 as reported above, a recent 162 study using machine learning to investigate adolescent SAs (with and without SI) showed that 163 the medial VMPFC was among the most discriminating regions, within a multivariate pattern 164 of fMRI brain activation in response to actively thinking about life- and death-related concepts, 165 for distinguishing between adolescent suicidal ideators with and without a history of attempt, 166 although in a very small sample size60. 167 168 Summary 169 Structural neuroimaging studies have consistently shown that alterations in the VLPFC and 170 VMPFC are implicated in SA across a range of mental disorders and age ranges. Reduced 171 VLPFC volumes were also associated with lethality of attempts, potentially mediated by 172 serotonergic dysfunction, although findings of serotonergic dysregulation remain inconsistent. 173 The involvement of structural and functional VLPFC and VMPFC alterations in SI remains 174 understudied. Viewing and regulating negative emotions and motivationally-valenced stimuli 175 has been linked to increased activation of the lateral and medial OFC in adult (including 176 young adults) SAs with MDD and BPD, and associated with poorer attempt planning and 177 higher impulsivity. Finally, higher VLPFC activity, including in the IFG, RLPFC and lateral 178 OFC, during cognitive control and response inhibition in relation to SA and SI in adults with 179 mood disorders has been reported across a number of studies. These increased activations 180 may, in the absence of task performance differences, reflect a need for greater engagement 181 of these regions for reaching similar performance in these individuals with STBs. In contrast, 182 adult SAs with SZ showed reduced activation in these regions during cognitive control, which 183 may suggest that there are some differences in the neural signatures of STBs between mood 184 and psychotic disorders, which will be an important direction for future study. 185 186 Dorsal Prefrontal Cortex (DPFC) 187 The DPFC can be broadly divided into dorsolateral (DLPFC) and dorsomedial PFC (DMPFC). 188 The DMPFC and the DLPFC together support top-down control of emotions and behaviors,17 189 cognitive flexibility and complex decision-making61. Deficits in these processes are thought to 190 have an important role in STBs, particularly in the transition from SI to behavior, as the 191 threshold to acting is lowered by decreased top-down behavioral inhibition, and diminished 192 flexibility in generating alternate and more adaptive behavioral choices15,62. Neuroimaging 193 evidence suggests that the DMPFC (medial portions of BA 8 and 9) is robustly recruited 194 during tasks that require mental state inference63,64. The DMPFC is further involved in 195 tracking decision conflict and reinforcement history65, as well as in emotion regulation66. The 196 DLPFC (BA46, lateral BA9) is involved in the conscious active control of planned behavior 197 and cognition, as well as working memory67. Access of the DLPFC to memory processing in 198 hippocampal regions is shared by the rostrolateral PFC (lateral frontal pole BA10), which has 199 been implicated in meta-cognitive awareness68,69. 200 8 201 DLPFC 202 Although the amount of evidence to date has been less than in VPFC, accumulating findings 203 also support a role for the DLPFC in STBs. Structural MRI studies support lower volume in 204 DLPFC in adult SAs across MDD32,33,70 and BD33,35. Studies showed lower DLPFC thickness 205 in adult SAs with MDD58 and SZ71, but have not examined cortical surface area. In addition, 206 lower DLPFC volume was associated with attempt lethality in mood and psychotic 207 disorders32,37. Higher baseline 5-HT1a receptor binding potential in the DLPFC was also 208 associated with higher lethality of future attempts and SI during a 2-year follow-up42. 209 210 Fuctionally, decreased lateral (and medial) DPFC when ten adults with self-reported 211 depression listened to their own narrative of their attempt was reported in a study in which 212 imaging was conducted close to the time of the attempts (one to four weeks prior72). A study 213 of adolescents with a history of SI showing lower right DLPFC activation during passive 214 viewing of negative emotional scenes, suggests that DLPFC decreases during processing of 215 negative emotional stimuli might be present early in the course of SI73. Higher right DLPFC 216 engagement was observed during regulation of responses to negative emotional scenes in 217 the same adolescents, suggesting the direction of DLPFC differences depends on the specific 218 task requirements (passive viewing versus regulating)73. Another study in adolescent SAs 219 with MDD suggests that the direction may also relate to the specific emotion, as passive 220 viewing of angry faces, but not happy faces, elicited higher DLPFC responses74, perhaps due 221 to the high sensitivity to criticism and social rejection previously reported in individuals with 222 STBs75,76. 223 224 The DLPFC’s critical role in decision-making in the context of evaluating the motivational 225 value of choices77 may be especially relevant to STBs. Blunted DLPFC activation was 226 observed when evaluating risky versus safe options in adult SAs with MDD47 and when 227 evaluating lower immediate rewards versus larger delayed rewards in older adults with MDD 228 and well-planned suicide attempts78. This is in line with a behavioral study of older adult SAs 229 with MDD showing lower levels of delay discounting (or impulsive decision-making) and better 230 planning in suicide attempts79. Increased DLPFC activation has also been observed in adults 231 with STBs across a range of “cold” cognitive control tasks, especially those requiring inhibition 232 of automatic response tendencies. These included paradigms such as continuous 233 performance, stop signal, go-no go and stroop tasks studied in adult SAs with MDD or BD 234 with psychotic features52, SZ80, and ideators with SZ81 or PTSD and MDD in veterans53. 235 Elevated activation in DLPFC was observed in suicidal ideators with past attempts compared 236 to ideators without attempts while performing a continuous performance task52. Single photon 237 emission tomography (SPECT) and PET studies showing lower resting regional cerebral 238 blood flow (rCBF) and glucose metabolic rates (rCMRglu) in the DLPFC in adult SAs with 239 9 mood disorders82,83. Moreover, lower DLPFC rCMRglu was associated with higher lethality of 240 attempt84 and with SI with a plan versus ideation without a specific plan85 in adults with MDD. 241 242 DMPFC 243 Structural MRI studies also support lower volume in DMPFC in adult SAs with MDD32,70 and 244 BD33,35. Although less studied than DLPFC in functional neuroimaging, lower DMPFC was 245 reported in the study in which adults with depression listened to their own narrative of their 246 recent attempt, which was especially pronounced during mental pain aspects of the 247 narrative72. Furthermore, decreased activation was also found in DMPFC during viewing of 248 angry faces in adult SAs with MDD48. 249 250 Summary 251 Structural alterations in both the DLPFC and DMPFC have been consistently observed in 252 adults across mental disorders and structural alterations in the DLPFC have been associated 253 with attempt lethality. This latter finding, together with findings of serotonergic dysfunction in 254 the DLPFC being associated with higher lethality of future attempts during a 2-year follow-up, 255 implicates DLPFC structure and serotonergic system functioning in STB risk. With regard to 256 DLPFC and DMPFC functioning, there is convergence in showing differences related to STBs 257 during processing of negative emotional stimuli, although the direction of effects (activation 258 increases versus decreases) in the DLPFC differed across studies, with contributions to the 259 differences unclear as the studies differed across multiple variables. Elevated activation while 260 performing cognitive control tasks (in the absence of performance differences) together with 261 lower resting regional cerebral blood flow and glucose metabolic rates in the DLPFC could 262 suggest that increased DPFC may be recruited for reaching similar task performance perhaps 263 due to lower baseline levels of activation in the DPFC regions. Findings from one study 264 suggests that functional DLPFC alterations during cognitive control can discriminate between 265 suicidal ideators with past attempts and ideators without attempts. Moreover, blunted DLPFC 266 activation when evaluating the value of different decision options may also represent a risk 267 marker for SA, and especially for well-planned attempts. Well-planned versus impulsive 268 suicide attempts have been suggested to be different phenotypes86. From papers reviewed, 269 there was a greater concentration of findings in the VPFC in impulsive SAs and in DPFC in 270 planful SAs. Therefore, we speculate that the relative ventral versus dorsal localization of the 271 PFC abnormalities may contribute to the differing phenotypes. 272 273 Anterior cingulate cortex (ACC) 274 The ventral ACC consists of BA25 and ventral BA32 sub- and pre-genual to the corpus 275 callosum and plays a critical role in valuation and control of autonomic viscero-sensory 276 signals, the modulation of physiological responses to stress and the appraisal of internal 277 feelings16. The dorsal ACC (dorsal BA24 and 32) plays an important role in the appraisal of 278 10 actions (and adaptively adjust behaviour as a consequence) and reward-based decision-279 making16. 280 281 Structural MRI studies support lower volume in both ventral and dorsal ACC in adult SAs with 282 MDD34,58 and BD35, and that was related to a higher number of attempts in adolescents with 283 BPD and MDD87 and to higher lethality of attempts in adults with BPD88 and psychotic BD37. 284 Lower dorsal ACC (dACC) rCMRglu was associated with higher lethality of attempt84. In 285 addition, lower 5-HT1a binding in the ACC was associated with interim SI during a 2-year 286 follow-up in adults with MDD and past attempts42. A few studies that investigated 287 neurotransmitter systems other than 5-HT implicated the ACC in relation to SI. For example, 288 a positive association was shown between SI and ACC monoamine oxidase-A (MAO-A) 289 density in adults with BPD89. A relation between SI and increased ACC neuroinflammation (as 290 assessed by translocator protein (TSPO) availability) was reported in adults with MDD90. 291 Furthermore, dACC GABA concentrations were lower in adult female SA+SI compared to 292 clinical controls without SA or SI, however, this effect was no longer significant after 293 correcting for age91. 294 295 Viewing of angry faces elicited higher dACC responses in adolescent SAs with MDD74, 296 perhaps due to the high sensitivity to criticism and social rejection previously reported in 297 individuals with STBs75,76. In contrast, decreased activation was found in ACC (ROI capturing 298 both ventral and dorsal ACC) during viewing of sad faces in adult SA with MDD47. With regard 299 to positive stimuli, blunted ventral ACC responses were found during the anticipation of 300 reward in adult92, including elderly93, SAs with MDD. Elevated responses in the ventral ACC 301 have also been reported in relation to positive stimuli. For example, higher activation in 302 ventral ACC was seen in response to happy facial expressions48 and in response to actual 303 winning47 (in contrast to blunted response during reward anticipation) in adult SAs with MDD. 304 305 Summary 306 The ACC has mostly been studied in relation to emotional processing. Although various 307 studies have observed dorsal and ventral ACC activation alterations in adolescents and 308 adults with MDD and SA, the direction of alterations seem to be complex and dependent on 309 task condition and stimulus type (positive versus negative). One could perhaps interpret the 310 findings of increased dorsal and ventral activation in response to angry faces and to positive 311 stimuli, together with blunted ventral ACC activation during reward anticipation, as negative 312 biases, as they may reflect reduced reward anticipation (anticipation phase) versus increased 313 activation in response to negative stimuli and in relation to positive prediction errors in 314 response to positive stimuli (outcome phase). The finding of a positive relation between SI 315 and ACC neuroinflammation, together with findings of increased inflammatory markers in the 316 ACC in postmortem studies of people who died by suicide48 and in blood and cerebrospinal 317 11 fluid in people with ideation and a history of violent or high intent attempts49,50, suggests that 318 neuroinflammation in the ACC may constitute a promising target for future studies of STBs. 319 320 Insula 321 The insular cortex is a key hub in emotional processing with connectivity to PFC, particularly 322 VPFC, as well as mesial temporal structures94. The insula plays an important role in 323 interoceptive awareness for positive and negative internal states95, including emotional and 324 other types of pain, and understanding and sharing of other people's emotional states96,97. 325 Only in more recent studies has insula structure and function and related-behavior been 326 investigated for its role in STBs. For example, on a behavioral level, interoceptive deficits 327 have been reported among SAs compared to individuals who only thought about or planned 328 suicide among general psychiatric outpatient adults98 and predicted SI severity at 6-month 329 follow-up in community adolescents99. 330 331 Smaller insula volume has been reported in adult SAs with BPD100, in a combined group with 332 SZ/SZA/psychotic BD37 and elderly with MDD70. Lower insula thickness was observed in 333 adults in relation to suicide attempts in SZ71 and SI in MDD38. Smaller insula volume was 334 associated with higher attempt lethality and lower impulsivity in BPD88,100. In contrast, larger 335 insula volumes were reported in relation to attempt lethality in adults with BD101. It is possible 336 that the type of insula differences relate to specific characteristics of the high lethality 337 attempters, since larger insula volumes were also found in association with higher lifetime 338 history of aggression in BPD88. Some findings in the PFC noted above in MDD extended to 339 the insula, including of associations between baseline 5-HT1a binding potential with SI and 340 lethality of future attempts within a 2-year follow-up period42 and of increased 341 neuroinflammation (TPSO availability)90. 342 343 SPECT research showed higher insula rCBF in adult SAs with MDD102 at rest and higher 344 insula fMRI activation was found in adults with MDD or BD with psychotic features during a 345 cognitive control task with insula activity related to higher intensity of SI52. Higher insula fMRI 346 activation was also associated with lower subjective value of gain and loss in adult SAs with 347 MDD92. Lower activation in the posterior insula during social exclusion was found in adult 348 SA’s with MDD or BD, which was suggested to indicate a higher tolerance to pain via 349 repeated exposure to painful and provocative experiences in subjects vulnerable for 350 suicide103. 351 352 Summary 353 Smaller insula volume has been associated with SAs and lower impulsivity in adults across 354 various mental disorders, whilst, both smaller and larger insula volumes have been 355 associated with higher attempt lethality. fMRI studies found higher insular activation related to 356 during reward processing and cognitive control in adult SA with MDD, while lower insula 357 12 activation was associated with a higher tolerance to social pain in adult SA’s with MDD or BD. 358 Thus, there is preliminary evidence for an involvement of insular structural and functional 359 alterations in SI and SAs. However, since very few studies have focussed on the insula and 360 both decreases and increases in these alterations have been reported, more research is 361 needed to elucidate the role of the insula in STBs. Interestingly, immune challenges activate 362 interoceptive brain pathways (including the insula), triggering alterations in mood and 363 cognition, motivation, and neurovegetative processes104. Together with preliminary evidence 364 of increased neuroinflammation in the insula related to SI, this suggests that the insula may 365 be an important region for future studies of neuroinflammation and STBs. 366 367 Amygdala and Hippocampus 368 Due to their roles in processing of emotion, emotional memory and the stress response105–108, 369 the mesial temporal amygdala, hippocampus and entorhinal cortex (BA28, within the adjacent 370 parahippocampal gyrus) are also thought to be involved in STBs. However, findings have 371 been inconsistent. Larger amygdala volumes were reported in adult SAs with MDD109 and 372 SZ110, but more studies have not detected significant amygdala findings33,38,100,101,111–113. 373 Smaller hippocampus volumes were reported in adult SAs with MDD114 and adolescent and 374 young adult SAs with BD36, and one study reported a smaller parahippocampal gyrus115. 375 However, more studies have not detected associations of hippocampus or parahippocampus 376 volume with STBs33–35,38,101,109–112. While difficulties detecting differences in these small mesial 377 temporal structures may relate to imaging methods, it may be that mesial temporal alterations 378 are only apparent in specific subgroups of people with STBs. For example, high lethality 379 attempts were associated with smaller volumes of the hippocampus and parahippocampal 380 gyrus in adult SAs with BPD88,100. In addition, amygdala and hippocampus volumes were 381 negatively associated with impulsivity in patients with low lethality attempts88, and amygdala 382 volume positively associated with self-aggression in SAs with SZ110. 383 384 PET studies have shown preliminary evidence for a role of serotonergic alterations in the 385 amygdala and hippocampus in STBs. Increased hippocampus 5-HT2a receptor binding and 386 5-HT release was observed in adult SAs with BPD116 and in adults with MDD and high 387 lethality attempts40 respectively, compared to HCs. Recently, baseline 5-HT1a receptor 388 binding in the amygdala, hippocampus and parahippocampal gyrus was associated with 389 higher SI during a 2-year follow-up in adults with MDD42. 390 391 Few fMRI studies have focused on mesial temporal ROIs. An activation study focusing on the 392 amygdala found no association between SI and amygdala functioning during emotion 393 processing in 10 ideators117. Increases were observed during autobiographical recall of 394 mental pain experienced during an ideator’s own attempt in right parahippocampal gyrus 395 versus suicide action in left hippocampus72. In contrast, parahippocampal gyrus activation 396 was blunted in adult SAs with MDD choosing between a smaller immediate reward versus a 397 13 larger but delayed reward, especially when the two rewards were more than 1 year apart78. 398 Given the role of the parahippocampal gyrus in prospection118, its blunted response to 399 prospects with longer versus shorter delays may represent a neural substrate of impaired 400 prospection in SAs119–121, potentially undermining the deterrents and alternative solutions 401 during a suicidal crisis. 402 403 Summary 404 Although structural alterations in the amygdala and the hippocampus have been consistently 405 implicated in mental disorders122–124, the majority of studies reviewed do not report structural 406 alterations in these regions in relation to STBs. These mixed findings could perhaps be 407 explained if additional involvement of the amygdala and hippocampus in STBs beyond their 408 role in mental disorders is subtle with small effects only apparent in studies with very large 409 sample sizes. This is consistent with a post-hoc power analysis based on observed effect 410 sizes in the largest study on subcortical volumes in STBs to date111. Alternatively, mesial 411 temporal structural alterations may only become apparent in specific subgroups of people 412 with STBs. Preliminary evidence suggests serotonergic alterations in the amygdala and 413 hippocampus linked to SA as well as SI across mental disorders, and altered functioning in 414 these regions in relation to increased autobiographical recall of mental pain, blunted 415 immediate reward processing and impaired prospection in patients with SA, although 416 molecular and functional studies focussing on these regions are still scarce. 417 418 Striatum and Thalamus 419 The ventral striatum includes the nucleus accumbens and ventral parts of the putamen and 420 caudate125 and is a core region of the reward network126. Dorsal striatum, including dorsal 421 caudate and putamen, functions include initiating action, inhibitory control, and stimulus-422 response learning127. The striatum projects to the frontal lobe via the thalamus125, which is 423 also involved in sensory processing128. 424 425 Lower caudate and putamen volumes have been reported in adult SAs with MDD in 426 comparison to MDD non-attempters34,129 and HCs130, and putamen volumes were negatively 427 associated with impulsivity129. Lower putamen binding of the serotonin transporter (5-HTT) 428 was also reported in adult SAs with MDD compared to HCs131, and was negatively associated 429 with impulsivity132. However, striatal 5-HT binding was positively associated with SI in adult 430 SAs with MDD133. With regard to the thalamus, higher volumes were reported in veterans with 431 traumatic brain injury and suicide attempts134 and higher 5-HT synthesis reported in adult SAs 432 with a mix of psychiatric diagnoses40. However, other studies, including the largest study to 433 date of individuals with STBS (N=451), have not detected associations of striatum and 434 thalamus volume with STBs33,38,111. 435 436 14 Few fMRI studies have examined striatum and thalamus regions. Positive correlations were 437 observed between intensity of past SI and dorsal striatum responses during cognitive control 438 in adults with MDD or BD with psychotic features52. Lower putamen activation in adults with 439 BD during a motor task was associated with higher SI135. Higher thalamus activation was 440 observed when viewing knives (versus landscapes)136 and higher thalamus activation during 441 response inhibition (go-no-go task) was associated with higher levels of mental pain and 442 suicide intent137 in adults with SAs and MDD. 443 444 Summary 445 Mixed findings were reported for the involvement of structural alterations in the striatum and 446 thalamus in relation to STBs, with the largest sample to date showing no associations in 447 people with MDD. Increased dorsal striatum responses were found during cognitive control in 448 the absence of performance differences in individuals with SI, suggesting greater 449 engagement of this region to reach similar levels of cognitive control. Higher thalamus 450 activations were reported during emotion processing and inhibition and associated with SI in 451 adult SA with MDD. Structural and 5-HTT alterations in the dorsal striatum specifically linked 452 to impulsivity in adult MDD with SA converge with findings of functional alterations in the 453 dorsal striatum in relation to diminished cognitive and affective control associated with SI. Of 454 note, alterations in the ventral striatum have been proposed to underlie reduced reward 455 anticipation and anhedonia in individuals with STBs138. No studies reported ventral striatal 456 activity, however, ventral striatal connectivity findings during reward processing and under 457 rest are discussed in the “Structural and Functional Connectivity” section below. 458 459 Posterior Structures 460 The temporal association cortices are involved in perceptual processing of faces and other 461 complex object features139,140, auditory information and language141. Consistent with this, 462 structural MRI findings in lateral temporal cortex were observed in adult SAs with 463 SZ/SZA/psychotic BD and other disorders, such as mood disorders, in which psychotic 464 misperceptions can be observed. Lower middle and superior temporal gyrus volume was 465 found in adult SAs with primary psychotic disorders37,142 and BPD100, and lower thickness of 466 middle and superior temporal gyri were observed in adult SAs with SZ71. Lower middle and 467 superior temporal volumes were also associated with high lethality attempts in adults SAs 468 with BPD88,100. Serotonin system studies have yielded various results including lower 5-HTT 469 temporal binding associated with higher impulsivity in adult SAs with different mental 470 disorders132, and higher baseline 5-HT1a temporal lobe receptor binding in adults SAs with 471 MDD42 associated with higher levels of SI at 2-year follow-up. Functional MRI studies also 472 suggest a role of the lateral temporal lobe in emotion processing in STBs. Specifically, 473 adolescent SAs with MDD showed enhanced right middle temporal gyrus activation during 474 passive viewing of angry, happy and neutral facial expressions74 and during recall and re-475 imagination of suicidal episodes in adult SAs with MDD72. SI was associated with increased 476 15 superior temporal activation during error processing in veterans with traumatic brain injury53. 477 In contrast, lower perfusion in these temporal regions during rest (measured by rCBF) was 478 reported in adults with MDD and SA82,143. 479 480 A few studies implicate other posterior brain regions including posterior cingulate cortex 481 (PCC) and cerebellum in STBs (Tables 1-3). The PCC is implicated in psychological 482 processes that may be linked to STBs, including controlling the vividness of negative mental 483 imagery144 and enhancing self-referential processing145. Lower PCC gray matter volume was 484 found in adult SAs with MDD146. Decreased PCC activation was observed during cognitive 485 control in adult SAs with psychotic mood disorders52 and during self-referential processing in 486 adolescent ideators with MDD147, although adult SAs with depressive disorders, compared to 487 HCs, showed increased PCC response when viewing knives136. 488 489 The cerebellum is increasingly recognized for its involvement in emotional processes148,149. 490 Lower volumes of the cerebellum were reported in adult and adolescent SAs with MDD or 491 BD36,70,146,150. Functionally, while adult SAs with MDD showed increased cerebellum activation 492 during recall and re-imagination of their own suicidal episode72 and while viewing angry faces, 493 they showed decreased activation while viewing happy faces48. Decreased activation was 494 also observed while passively viewing negative emotional pictures in adolescents with a 495 history of SI73. Finally, ketamine-induced reductions in SI were associated with increases in 496 rCMRglu including in cerebellum151. 497 498 Summary 499 Lower middle and superior temporal gyri volumes have been reported in 6 studies across a 500 range of mental disorders, and related to high lethality attempts and higher impulsivity. 501 Serotonergic alterations in these regions have not been extensively investigated and 502 directions of reported effects are mixed. Increased activation in middle and superior temporal 503 gyri have also been reported in adults and adolescents with SA and MDD, especially in 504 relation to emotion processing. Preliminary evidence suggests a role for the PCC and 505 cerebellum in STBs, especially in relation to self-referential, and (autobiographical) emotion 506 processing, but studies investigating these regions remain scarce. Of interest, one study 507 suggests that ketamine-induced changes in SI are associated with ketmaine-induced 508 increased in cerebellum rCMRglu. This medication-related finding is a potential lead in 509 understanding brain mechanisms that may be helpful targets for suicide prevention 510 interventions, but requires replication. 511 512 Structural and Functional Connectivity 513 Disturbances in the structure and function within brain regions can result in alterations in brain 514 networks, including the ability of brain regions to coordinate their activity in a system. System 515 dysfunction can also result from abnormalities in the connections between regions. 516 16 Increasingly, abnormalities in the structural and functional connections between brain regions 517 within larger-scale brain networks have been reported in studies of STBs. 518 519 Connections of the medial VPFC with other cortical midline structures (PCC, precuneus) and 520 temporal and parietal regions are implicated in the brain’s default mode network and play an 521 important role in self-referential processes, social cognition, autobiographical memory and 522 prospective imaging118. Lower resting functional connectivity was reported in the default mode 523 network in adolescents with MDD and SI152. These findings are in line with findings of lower 524 resting functional connectivity of rostral ACC with medial OFC, precuneus and temporal pole 525 in adults with MDD and SI153 and within the precuneus in young adult SAs with MDD154. 526 Structurally, lower fractional anisotropy (FA; thought to reflect the structural integrity of white 527 matter and the neuronal connections it contains155) in the medial VPFC was reported in adult 528 SAs with BD, which was associated with higher motor impulsivity156. Lower FA was also 529 reported in the ventral cingulum (connecting posterior and temporal default mode network 530 regions) in adults with MDD and SI38. Lower FA of the corpus callosum genu, that provides 531 connections of interhemispheric anterior default mode network regions, was associated with a 532 higher number of SAs in BD, MDD and BPD157,158. 533 534 The limbic network includes the amygdala, medial and lateral OFC, medial temporal regions, 535 thalamus and basal ganglia and is involved in emotional and autonomic processes159. Using 536 task fMRI, lower amygdala-medial VPFC/rostral PFC connectivity was found in adolescent 537 and young adult BD SAs while viewing happy and neutral facial expression, associated with 538 higher lethality of attempts and current SI36. During rest, greater amygdala connectivity with 539 lateral OFC, insula and middle temporal gyrus was found in adult SAs with MDD, with greater 540 amygdala-parahippocampus connectivity associated with SI160. These functional connectivity 541 alterations are in line with lower FA in the uncinate fasciculus, that provides major amygdala 542 connections, in adolescents SAs with BD36. 543 544 The medial OFC together with ventral striatum and the ventral tegmental area form core hubs 545 of a reward network, with additional limbic regions, DLPFC and dACC forming a wider reward 546 network subserving reward-related memory and evaluation126. An fMRI study investigating 547 connectivity during reward processing showed a positive correlation between SI and 548 connectivity of the left ventral striatum with dACC, DMPFC and DLPFC during loss trials in 549 adults with MDD161. Using resting state fMRI, Kim et al.162 found reduced connectivity in 550 circuitry resembling this reward network, including the OFC, striatum and thalamus, in adults 551 with MDD and recent (past month) SI. This is in line with findings of lower structural 552 connectivity between VPFC/OFC and striatal regions in adults with MDD and SI (33% also 553 had prior SA)163, and between the ACC and OFC (as measured by graph theory) in adult SAs 554 with MDD164. Lower FA was also reported in the anterior limb of the internal capsule 555 17 (connecting striatal and thalamic regions with the PFC) in adults with MDD and 556 attempts165,166. 557 558 The left and right DLPFC and DMPFC together with parietal regions comprise a network key 559 in cognitive control of thought, emotion and behavior (executive control network167). Lower 560 executive control network coherence during rest has been associated with both lifetime SI 561 and past attempts in adolescents with MDD152, in line with findings of lower DLPFC resting 562 state connectivity in young adult SAs with MDD associated with higher impulsivity168 and 563 reduced white matter integrity (FA) in the DMPFC in adult SAs with MDD169. 564 565 Alterations in dACC connectivity have been linked to STBs in the context of conflict-566 monitoring, with different patterns of dACC connectivity associated with SI versus SA in 567 adults with recent-onset SZ170. That is, the presence of lifetime SI was positively associated 568 with magnitude of functional connectivity of dACC with the precuneus, a core hub of the 569 default mode network. This may suggest a reduced capacity of the dACC to ‘switch off’ 570 default mode network activity associated with more internally focused attention, when 571 activation of externally focused cognitive processing is required170. In contrast, history of 572 suicide attempts was negatively associated with dACC connectivity with DPFC (BA9, 8, 573 lateral BA10), lateral VPFC (BA45), PCC, parietal regions (BA7,40) and superior and middle 574 temporal gyri (BA22, 39, 40)170. These findings may suggest that SI and SA have divergent 575 bases in dACC connectivity with default mode network versus lateral PFC circuitries, 576 respectively, in the context of monitoring conflict in adult SAs with recent-onset SZ. This is 577 supported by findings of abnormal conflict-related dACC connectivity with VLPFC, OFC, 578 insula and striatum associated with SI intensity, but altered dACC connectivity with DLPFC 579 and frontal motor regions associated with past suicide attempts in adults with MDD or BD with 580 psychotic features171. In addition, decreased functional connectivity of the dACC with bilateral 581 insula while viewing angry faces was also reported in adolescent SAs with MDD74. 582 Connectivity between the dACC and insula plays an important role in detecting salient internal 583 and external stimuli to guide behavior172 and has been implicated in the anticipation of 584 aversive experiences, especially in depressed individuals173. Reduced connectivity between 585 the dACC and bilateral insula may indicate inefficient strategies to process the salience of, 586 and select contextually appropriate behavioral responses to, negative emotional stimuli. 587 588 Summary 589 Emerging evidence suggests that resting state functional connectivity in the default mode 590 network, and white matter tracts connecting regions within this network, play a role in both SI 591 and SA across mental disorders. Functional and structural connectivity alterations in the 592 affective network have also been associated with SI and SA, as well as lethality of attempts, 593 both during emotion processing and during rest. Connectivity abnormalities in the reward 594 network have mostly been examined in adults with MDD, and both structural and functional 595 18 connectivity changes in regions of this network have been associated with SI and SA in this 596 group. Connectivity changes within and between regions implied in the cognitive control 597 network has been less extensively studied in relation to STBs, and the few studies conducted 598 suggest a role of lower resting state functional connectivity within this network in STBs in 599 adolescents and young adults with MDD. In addition, functional connectivity of dACC is 600 implicated in STBs, but with divergent connectivity patterns related to SI versus SA. 601 602 603 4. DISCUSSION 604 While the literature primarily includes cross-sectional studies with small sample sizes, 605 differing clinical populations and a wide range of imaging methods, there is emerging 606 convergence in the brain regions implicated in STBs. Taken together with the recent 607 increased momentum in studies on STBs (Figure 1), it is very hopeful that significant 608 advances in our understanding of brain mechanisms contributing to STBs are on the horizon. 609 A critical frontier is to identify markers for elevated risk, especially short term risk in the 610 transition from SI to attempt. While the majority of studies to date were on SAs, some studies 611 specifically investigated associations with SI, and brain regions found to be associated with 612 STBs have known roles in processes thought to contribute to STBs. Below we briefly 613 summarize the most convergent findings from the literature reviewed above, and propose 614 directions for future neuroimaging studies on the neurobiology of STB. 615 616 A Tentative Brain Model of STBs 617 Figure 3 summarizes convergent findings emerging from the reviewed literature on brain 618 alterations associated with STBs. Abnormalities in an extended VPFC system, including 619 regions of the default mode, affective and reward networks such as the ventral and rostral 620 ACC, insula, medial and lateral OFC, mesial temporal regions, ventral striatum and posterior 621 structures (lateral temporal, PCC, precuneus, cerebellum), and in the connections among 622 these regions, may be important in the excessive negative and blunted positive internal states 623 that can stimulate SI. This is in line with the well-established role of this extended VPFC 624 system in functions implicated in SI including appraisal of internally-generated emotions, self-625 referential processing, recall of emotional episodic memories, imagining future positive and 626 negative events, valuation of rewards, and integrating environmental stimuli to modulate 627 subjective emotional states16,17,67,118,126,148. A more lateral and dorsal system, including 628 DMPFC, DLPFC and dACC, and together with the IFG, may facilitate suicide behaviors 629 through their role in cognitive control of thought, emotion and behavior as well as cognitive 630 flexibility, complex decision-making (e.g. valuation of different decision options) and 631 planning16,17. A combination of VPFC and DPFC/IFG system disturbances may lead to very 632 high risk circumstances in which SI may convert to lethal actions via decreased top-down 633 inhibition of behavior or maladaptive, and inflexible decision-making and planning. 634 635 19 Alterations in the connections between these systems may contribute to the transition from 636 suicidal thoughts to behaviors. For example, the dACC has connections to both the 637 “emotional” ventral limbic system and the “cognitive” dorsal prefrontal system174. Consistent 638 with this, findings suggest differential connectivity of the dACC in relation to SI versus a 639 history of attempt, with dACC connectivity with VPFC, insula and striatal regions primarily 640 associated with SI and dACC connectivity with DPFC regions associated with suicide 641 attempts. The insula is also implicated in both SI and attempts and may perhaps be important 642 for the transition from SI to attempt. Although the involvement of the insula in STBs has had 643 little direct research focus, its critical involvement in interoceptive processing, detecting 644 salient internal and external stimuli, experiencing emotion and self-awareness175–177 suggests 645 an important role of the insula in SI. Additionally, the insula is implicated in disconnection from 646 bodily experiences, which in turn may lower the threshold to engaging in behaviors that harm 647 the body (in line with the acquired capability theory178) thus suggesting a role of the insula in 648 suicide behaviors. In line with important roles of dACC and insula circuitry (as part of the 649 ‘salience network’) in mediating or switching between the extended medial VPFC (default 650 mode, affective, reward) systems and the DPFC/IFG (executive control) system179–181, the 651 dACC and insula may represent integral hubs that facilitate the transition from SI to attempt. 652 However, this suggestion of the dACC and insula’s ability of mediating dynamic interactions 653 between the medial VPFC and DPFC/IFG systems and through these interactions playing a 654 role in the transition from SI to attempt remains highly speculative and will need to be 655 confirmed in future, preferably longitudinal, studies on the neurobiology of STB. 656 657 Future directions 658 The literature reviewed underscores the need for future studies to include larger sample sizes 659 and careful attention to developmental stages. In addition, studies employing longitudinal 660 designs are critically needed to identify risk markers for future suicide attempts in order to 661 develop improved preventive strategies. Recent preliminary evidence, from a rare longitudinal 662 study of adolescents and young adults with mood disorders over about 3 years, showed that 663 those with future attempts had lower VPFC volume and decreased FA in VPFC and DPFC 664 connections182, suggesting that these may be potential predictors for STBs already present in 665 adolescence. Longitudinal study over short time intervals are largely absent from the literature 666 and are critically-needed to assess proximal suicide risk. 667 668 Furthermore, studies focusing on identifying brain alterations that predict or fluctuate with 669 changes in STBs following treatment could provide urgently needed biomarkers for response 670 of STBs to existing interventions and could help develop novel treatments specifically 671 targeting these biomarkers. Preliminary evidence has implicated brain circuitry that may 672 mediate the reduction of suicide risk by treatments such as lithium and ketamine183–185. For 673 example, ketamine-induced reductions in SI were associated with increases in baseline 674 rCMRglu in a cluster including the cerebellum and occipital cortex151. In addition, in 57 adults 675 20 with BD with and without a history of attempt, lowest DLPFC, OFC, ACC, superior temporal 676 cortex, parietal and occipital cortex volumes were observed in SAs off-lithium, followed by 677 SAs and non-attempters on lithium, with the largest volumes in people with non-attempting 678 BD patients on lithium35. However, the study of other pharmacological and non-679 pharmacological treatments that can directly (e.g. deep brain stimulation) or indirectly target 680 the involved circuitry is needed. 681 682 Although adolescents have been less studied than adults, some findings have been similar. 683 For instance, alterations in structure of the VMPFC, VLPFC, DLPFC, DMPFC, ACC, lateral 684 temporal regions and parahippocampal gyrus were observed in both adolescents and adults 685 with SA. Furthermore, lower connectivity in regions related to the default mode network 686 during rest have been consistently reported across adolescents, young adults and adults with 687 SI and/or SA. Greater DLPFC activity in response to angry faces was also observed in both 688 adolescents and adults with STBs. The highly limited number of studies in adolescents and a 689 lack of different life stages in single studies prevents drawing conclusions about the overlap in 690 functional brain alterations across different stages of life and brain maturation. Thus, there is 691 some converging evidence across adolescent and adult samples, however, not all adult 692 findings have been observed in adolescents (see Tables 1-3 for a complete overview). This 693 may at least in part be due to the continued maturation of involved brain systems so that not 694 all features may be expressed until adulthood186,187. 695 696 Most studies only included a single disorder, impeding conclusions around shared versus 697 unique neural substrates of STBs across different mental disorders. Different studies 698 employed different in- and exclusion criteria, imaging methods and STB assessments. 699 Nonetheless, gray matter alterations in the VLPFC, DLPFC, ACC, and insula have been 700 consistently reported across the diagnostic categories reviewed (i.e., MDD, BPD, BD, SZ), 701 while lateral temporal alterations were more uniquely observed in psychotic disorders and 702 BPD. Reduced white matter integrity in ventral PFC regions was reported in relation to STBs 703 in both BD and MDD and lower FA in the corpus callosum was associated with a higher 704 number of attempts across BD, MDD and BPD. Functional brain alterations in relation to 705 emotion processing and regulation was investigated only in MDD, BD and BPD, with 706 consistent findings of increased VLPFC activation in response to negative emotional stimuli 707 across MDD and BPD, while the only study of adolescents with BD focused on amygdala-708 PFC connectivity36. Higher DLPFC activation during cognitive control tasks, in the absence of 709 task performance differences, was also consistently reported across MDD, BD, SZ and 710 PTSD. In contrast, higher versus lower IFG activation during cognitive control was observed 711 in mood disorders versus SZ, respectively. Alterations in dACC connectivity with default 712 mode, affective and reward network related regions in relation to SI, and alterations in dACC 713 connectivity with DLPFC in relation to SA during cognitive control, has been observed across 714 MDD, BD and SZ. Other functional domains such as reward processing, decision making, 715 21 social exclusion and self-referential processing, as well as resting state fMRI, have only been 716 investigated in a single disorder, i.e. depression. In one of the few studies to assess SAs 717 across MDD and BD (published subsequent to our literature search), VPFC gray matter 718 volume reductions and uncinate fasciculus FA reductions were common to both disorders, 719 suggesting that they may be important in risk for attempts across mood disorders188. 720 Preliminary findings of that study also indicated there may be differences in involved regions 721 in SAs between the disorders, with greater uncinate involvement in attempters in BD and 722 dorsal frontal white matter in MDD. These findings suggest both transdiagnostic and unique 723 gray and white matter targets for suicide prevention across mental disorders. 724 725 The examination of sex differences in STBs is a major gap in the neuroimaging literature and 726 a critical area for study. There are well-established sex-dependent features known in STBs189, 727 such as a higher rate of attempts in females and a higher lethality of attempts in males190–192. 728 Moreover, increased death rates by suicide between 1999 and 2017 as reported by the CDC 729 showed the rate of increase was substantially higher for females than males (53% and 26% 730 respectively). Female suicide rate increases were particularly high in youth/early adulthood 731 (ages 10-24) and middle age (45-64 years), when females have their highest risk, while the 732 highest risk for males is age 75 and older193. Of the studies in this review, 11 (8.4%) had 733 exclusively female and 15 (11.5%) exclusively male samples. Authors of six studies (4.6%) 734 commented on this homogeneity as a limitation, and 14 (10.7%) provided a rationale for 735 single sex studies. These included the need to avoid potential confounds of previously 736 reported sex-based effects e.g., previous reports of sex differences in cortical responses with 737 similar fMRI tasks (n=5) and corpus callosum structure (n=2); high male representation in the 738 group under investigation, such as military veterans (n=3); female-skewed prevalence of BPD 739 (n=1); males and females imaged on different scanners (n=1); small number of male SAs 740 (n=1); and an attempt to replicate previous work (n=1). Ninety studies (68.7%) included sex 741 as a covariate, controlling for potential differences. A small percentage of studies (13.8%) 742 included sex as a variable to assess potential interactions with STBs. Five studies (3.8%) 743 reported sex-related findings, though no relationship with STBs37,147,194–197 and one study 744 (0.8%) reported more females than males studied were SAs but reported no neuroimaging-745 related findings198. Chase and colleagues199, in a study that controlled for sex, noted that one 746 participant identified as transmale. Careful consideration is warranted in how sex and gender 747 are evaluated and categorized for analyses, including the importance of allowing subjects to 748 identify by gender and this self-identification of gender is considered in all studies. This is 749 particularly important because transgender and sexual-minority individuals are at increased 750 risk for STBs and death by suicide200,201. Collectively, these STB neuroimaging research data 751 highlight the urgent need for future work on sex and gender. 752 753 Integration of neuroimaging research across differential mechanistic levels including genetic, 754 molecular, social, and environmental risk factors will be crucial to the elucidation of a holistic 755 22 STB pathophysiology and mechanisms associated with vulnerability and resilience and thus 756 tailored intervention development. Sophisticated analytic methods, such as machine learning 757 techniques60, can be utilized to allow imaging risk markers to be identified at the level of the 758 individual instead of at a group level, a key ingredient for clinically viable biomarkers202 and 759 precision medicine. In addition, the use of high resolution ultra-high field strength MRI 760 methods and more specific functional neuroimaging tasks may further enhance ability to 761 parse roles of specific brain regions, such as PFC subregions203. Molecular imaging has 762 produced important leads, such as in serotonergic and inflammation mechanisms (Table 2), 763 consistent with post-mortem, genetic and peripheral biomarker studies implicating an 764 important role of the serotonergic system and inflammatory mechanisms in STBs204,205. 765 Imaging of molecular mechanisms other than serotonin remains scarce, with for example only 766 four magnetic resonance spectroscopy (1H-MRS) studies examining glutamatergic and 767 Gamma-Aminobutyric Acid (GABA)-ergic mechanisms that have been published to date 768 (Table 2). Given post-mortem findings of altered glutamatergic and GABA-ergic gene 769 expression and receptor availability in suicide victims mainly in the prefrontal cortex, ACC and 770 hippocampus206–208, future 1H-MRS should clarify the role of these neurotransmitters in STBs 771 in vivo. In addition, a generation of new methods to identify other currently implicated and 772 novel molecular mechanisms is needed. For example, a link between oxytocin and SI was 773 recently suggested209. 774 775 Future studies should also further investigate the more elusive subjective aspects of suicide 776 risk, such as SI, as well as implicated psychological constructs such as hopelessness, 777 rumination and anhedonia210–212, which may be relevant across diagnoses. Similarly, social 778 experiences such as childhood maltreatment and peer bullying form a strong prelude to STB 779 in later life209,213,214, and impact on the neural structures implicated in STB (e.g. DMPFC 780 structure and function215,216). Therefore, adverse experiences should be taken into account in 781 future studies on the neurobiology of STB. 782 783 Links identified between neuroimaging measures and behaviors outside of the scanner could 784 facilitate the development of less invasive and more easily and widely disseminated risk 785 detection methods. Furthermore, investigations of larger samples can be facilitated by 786 international collaborations that pool existing data across many different samples, such as the 787 MQ HOPES (https://www.mqmentalhealth.org/research/profiles/overcome-and-predict-the-788 emergence-of-suicide) and ENIGMA STB (http://enigma.ini.usc.edu/ongoing/enigma-stb/) 789 consortia. These initiatives represent cost-effective ways to substantially increase statistical 790 power, which could provide more robust and reliable findings217, and ability to study age, 791 gender and sex effects, as well as unique and shared mechanisms associated with STBs 792 across different mental disorders. Furthermore, large combined datasets will provide a unique 793 opportunity to identify and test reproducibility of different pathways to suicide that may differ 794 across many parameters including a range from the impulsive to the highly planned. Such 795 23 efforts will need to take into account the variety of assessment methods used for STBs across 796 studies. One way to address with this important challenge is through the examination of the 797 presence versus absence of suicidal behavior (attempts), and/or suicidal ideation (with or 798 without intent and/or a plan) across assessment types (see for example Renteria et al.111). 799 Another approach could involve standardization of scores across different instruments by 800 developing a common metric218. Standardization of STB assessment across future studies 801 would significantly facilitate the sharing of data, and thereby advancing our understanding of 802 brain-based STB vulnerability. 803 804 Conclusions 805 More than two decades of neuroimaging studies on STBs suggests a transdiagnostic model 806 for STBs in which an extended VPFC system may be important in the excessive negative and 807 blunted positive internal states that can stimulate SI and a DPFC/IFG system that may 808 facilitate suicide attempt behaviors. Interactions between these systems are likely important in 809 the transition from ideation to attempt, perhaps mediated by dACC and insula regions, but 810 require further investigation. With the exponential growth of research on STBs, including the 811 initiation of large global efforts, it is hopeful that suicide prevention will soon be more 812 effectively targeted, reducing the tragic loss of life to suicide. 813 1 Acknowledgments This work was supported by the MQ Brighter Futures Award MQBFC/2 (ALvH, HB, LS), by the National Institute of Mental Health of the National Institutes of Health under Award Number R01MH117601 (LS), RC1MH088366 (HPB), R01MH113230 (HPB), R61MH111929 (HPB), T32MH014276 (ETCL), and T32DA022975 (ETCL). LS is supported by a NHMRC Career Development Fellowship (1140764). ALVH is supported by a Royal Society Dorothy Hodgkin Fellowship (DH15017). LAA is supported by the American Foundation for Suicide Prevention, Brain and Behavior Foundation/NARSAD, Robert E. Leet and Clara M. Guthrie Patterson Trust, and Department of Veterans Affairs. HPB is supported by the American Foundation for Suicide Prevention and For the Love of Travis Foundation. Conflicts of interest The authors declare no competing financial interests. HPB received an honorarium for a talk at Aetna. 2 REFERENCES 1 World Health Organization. Public health action for the prevention of suicide: A framework. Geneva, Switzerland, 2012. 2 Substance Abuse and Mental Health Services Administration, Results from the 2013 National Survey on Drug Use and Health: Mental Health Findings, NSDUH Series H- 49, HHS Publication No. (SMA) 14-4887. Rockville, MD, 2014. 3 Eaton D, Kann L, Kinchen S, Shanklin S, Flint K, Hawkins J et al. Youth risk behavior surveillance–United States, 2011. Morbidity and Mortality Weekly Report. Surveillance Summaries. Surveill Summ 2012; 61: 1–162. 4 Curtin SC, Warner M, Hedegaard H. NCHS Data Brief No. 241: Increase in Suicide in the United States, 1999 – 2014 Key findings. 2016.http://www.cdc.gov/nchs/products/databriefs/db241.htm. 5 Franklin JC, Ribeiro JD, Fox KR, Bentley KH, Kleiman EM, Huang X et al. Risk factors for suicidal thoughts and behaviors: A meta-analysis of 50 years of research. Psychol Bull 2017; 143: 187–232. 6 Balcioglu YH, Kose S. Neural substrates of suicide and suicidal behaviour: from a neuroimaging perspective. Psychiatry Clin Psychopharmacol 2018; 28: 314–328. 7 Bani-Fatemi A, Tasmim S, Graff-Guerrero A, Gerretsen P, Strauss J, Kolla N et al. Structural and functional alterations of the suicidal brain: An updated review of neuroimaging studies. Psychiatry Res - Neuroimaging 2018; 278: 77–91. 8 Domínguez-Baleón C, Gutiérrez-Mondragón LF, Campos-González AI, Rentería ME. Neuroimaging studies of suicidal behavior and non-suicidal self-injury in psychiatric patients: A systematic review. Front Psychiatry 2018; 9. doi:10.3389/fpsyt.2018.00500. 9 Jollant F, Lawrence NL, Olié E, Guillaume S, Courtet P. The suicidal mind and brain: A review of neuropsychological and neuroimaging studies. World J Biol Psychiatry 2011; 12: 319–339. 10 van Heeringen K, Bijttebier S, Desmyter S, Vervaet M, Baeken C. Is there a neuroanatomical basis of the vulnerability to suicidal behavior? A coordinate-based meta-analysis of structural and functional MRI studies. Front Hum Neurosci 2014; 8: 1–8. 11 Hamza CA, Stewart SL, Willoughby T. Examining the link between nonsuicidal self- injury and suicidal behavior: A review of the literature and an integrated model. Clin Psychol Rev 2012; 32: 482–495. 12 Chesney E, Goodwin GM, Fazel S. Risks of all-cause and suicide mortality in mental disorders: A meta-review. World Psychiatry 2014; 13: 153–160. 13 Turecki G, Brent DA. Suicide and suicidal behaviour. Lancet 2016; 387: 1227–1239. 14 Hawton K, van Heeringen K. Suicide. Lancet 2009; 373: 1372–1381. 15 Dombrovski AY, Hallquist MN. The decision neuroscience perspective on suicidal behavior: Evidence and hypotheses. Curr Opin Psychiatry 2017; 30: 7–14. 3 16 Dixon ML, Thiruchselvam R, Todd R, Christoff K. Emotion and the prefrontal cortex: An integrative review. Psychol Bull 2017; 143: 1033–1081. 17 Mitchell DGV. The nexus between decision making and emotion regulation: A review of convergent neurocognitive substrates. Behav Brain Res 2011; 217: 215–231. 18 Murray RJ, Schaer M, Debbané M. Degrees of separation: A quantitative neuroimaging meta-analysis investigating self-specificity and shared neural activation between self- and other-reflection. Neurosci Biobehav Rev 2012; 36: 1043–1059. 19 Bechara A, Damasio AR. The somatic marker hypothesis: A neural theory of economic decision. Games Econ Behav 2005; 52: 336–372. 20 D’Argembeau A, Van der Linden M. Predicting the phenomenology of episodic future thoughts. Conscious Cogn 2012; 21: 1198–1206. 21 Levy DJ, Glimcher PW. The root of all value: a neural common currency for choice. Curr Opin Neurobiol 2012; 22: 1027–1038. 22 Urry HL, Reekum CM van, Johnstone T, Kalin NH, Thurow ME, Schaefer HS et al. Amygdala and Ventromedial Prefrontal Cortex Are Inversely Coupled during Regulation of Negative Affect and Predict the Diurnal Pattern of Cortisol Secretion among Older Adults. J Neurosci 2006; 26: 4415–4425. 23 Mitchell JP, Banaji MR, Macrae CN. The link between social cognition and self- referential thought in the medial prefrontal cortex. J Cogn Neurosci 2005; 17: 1306– 1315. 24 Drevets WC. Neuroimaging and neuropathological studies of depression: Implications for the cognitive-emotional features of mood disorders. Curr Opin Neurobiol 2001; 11: 240–249. 25 Jollant F, Lawrence NL, Olié E, Guillaume S, Courtet P. The suicidal mind and brain: A review of neuropsychological and neuroimaging studies. World J Biol Psychiatry 2011; 12: 319–339. 26 Bratman GN, Hamilton JP, Hahn KS, Daily GC, Gross JJ. Nature experience reduces rumination and subgenual prefrontal cortex activation. Proc Natl Acad Sci 2015; 112: 8567–8572. 27 Bertossi E, Ciaramelli E. Ventromedial prefrontal damage reduces mind-wandering and biases its temporal focus. Soc Cogn Affect Neurosci 2016; 11: 1783–1791. 28 Burke TA, Connolly SL, Hamilton JL, Stange JP, Lyn Y, Alloy LB. Cognitive Risk and Protective Factors for Suicidal Ideation: A Two Year Longitudinal Study in Adolescence. J Abnorm Child Psychol 2016; 44: 1145–1160. 29 Ducasse D, Loas G, Dassa D, Gramaglia C, Zeppegno P, Guillaume S et al. Anhedonia is associated with suicidal ideation independently of depression: A meta- analysis. Depress Anxiety 2018; 35: 382–392. 30 Miranda R, Valderrama J, Tsypes A, Gadol E, Gallagher M. Cognitive inflexibility and suicidal ideation: Mediating role of brooding and hopelessness. Psychiatry Res 2013; 210: 174–181. 4 31 Smith JM, Alloy LB, Abramson LY. Cognitive Vulnerability to Depression, Rumination, Hopelessness, and Suicidal Ideation: Multiple Pathways to Self-Injurious Thinking. Suicide Life-Threatening Behav 2006; 36: 443–454. 32 Ding Y, Lawrence N, Olié E, Cyprien F, Le Bars E, Bonafé A et al. Prefrontal cortex markers of suicidal vulnerability in mood disorders: A model-based structural neuroimaging study with a translational perspective. Transl Psychiatry 2015; 5: e516. 33 Gosnell SN, Velasquez KM, Molfese DL, Molfese PJ, Madan A, Fowler JC et al. Prefrontal cortex, temporal cortex, and hippocampus volume are affected in suicidal psychiatric patients. Psychiatry Res - Neuroimaging 2016; 256: 50–56. 34 Wagner G, Koch K, Schachtzabel C, Schultz CC, Sauer H, Schlösser RG. Structural brain alterations in patients with major depressive disorder and high risk for suicide: Evidence for a distinct neurobiological entity? Neuroimage 2011; 54: 1607–1614. 35 Benedetti F, Radaelli D, Poletti S, Locatelli C, Falini A, Colombo C et al. Opposite effects of suicidality and lithium on gray matter volumes in bipolar depression. J Affect Disord 2011; 135: 139–147. 36 Johnston JAY, Wang F, Liu JJ, Blond BN, Wallace A, Liu JJ et al. Multimodal neuroimaging of frontolimbic structure and function associated with suicide attempts in adolescents and young adults with bipolar disorder. Am J Psychiatry 2017; 174: 667– 675. 37 Giakoumatos CI, Mathew IT. Are Structural Brain Abnormalities Associated With Suicidal Behaviour in Patients With Psiychotic Disorder? 2014; 47: 1389–1395. 38 Taylor WD, Boyd B, McQuoid DR, Kudra K, Saleh A, MacFall JR. Widespread white matter but focal gray matter alterations in depressed individuals with thoughts of death. Prog Neuro-Psychopharmacology Biol Psychiatry 2015; 62: 22–28. 39 Arango V, Underwood MD, Mann JJ. Serotonin brain circuits involved in major depression and suicide. Prog Brain Res 2002; 136: 443–453. 40 Leyton M, Paquette V, Gravel P, Rosa-Neto P, Weston F, Diksic M et al. α- [11C]methyl-L-tryptophan trapping in the orbital and ventral medial prefrontal cortex of suicide attempters. Eur Neuropsychopharmacol 2006; 16: 220–223. 41 Sullivan GM, Oquendo MA, Milak M, Miller JM, Burke A, Ogden RT et al. Positron emission tomography quantification of serotonin1A receptor binding in suicide attempters with major depressive disorder. JAMA Psychiatry 2015; 72: 169–178. 42 Oquendo MA, Galfalvy H, Sullivan GM, Miller JM, Milak MM, Elizabeth Sublette M et al. Positron emission tomographic imaging of the serotonergic system and prediction of risk and lethality of future suicidal behavior. JAMA Psychiatry 2016; 73: 1048–1055. 43 Parsey R V., Oquendo MA, Ogden RT, Olvet DM, Simpson N, Huang YY et al. Altered serotonin 1A binding in major depression: A [carbonyl-C-11] WAY100635 positron emission tomography study. Biol Psychiatry 2006; 59: 106–113. 44 Miller JM, Everett BA, Oquendo MA, Ogden RT, Mann JJ, Parsey R V. Positron Emission Tomography Quantification of Serotonin Transporter Binding in Medication- 5 Free Bipolar Disorder. Synapse 2016; 70: 24–32. 45 Mann JJ, Metts A V., Ogden RT, Mathis CA, Rubin-Falcone H, Gong Z et al. Quantification of 5-HT1Aand 5-HT2Areceptor Binding in Depressed Suicide Attempters and Non-Attempters. Arch Suicide Res 2018; 1118: 1–12. 46 Cannon DM, Ichise M, Fromm SJ, Nugent AC, Rollis D, Gandhi SK et al. Serotonin Transporter Binding in Bipolar Disorder Assessed using [11C]DASB and Positron Emission Tomography. Biol Psychiatry 2006; 60: 207–217. 47 Olié E, Ding Y, Le Bars E, de Champfleur NM, Mura T, Bonafé A et al. Processing of decision-making and social threat in patients with history of suicidal attempt: A neuroimaging replication study. Psychiatry Res - Neuroimaging 2015; 234: 369–377. 48 Jollant F, Lawrence NS, Giampietro V, Brammer MJ, Fullana MA, Drapier D et al. Orbitofrontal cortex response to angry faces in men with histories of suicide attempts. Am J Psychiatry 2008; 165: 740–748. 49 Vanyukov PM, Szanto K, Siegle GJ, Hallquist MN, Reynolds CF, Aizenstein HJ et al. Impulsive traits and unplanned suicide attempts predict exaggerated prefrontal response to angry faces in the elderly. Am J Geriatr Psychiatry 2015; 23: 829–839. 50 Silvers JA, Hubbard AD, Chaudhury S, Biggs E, Shu J, Grunebaum MF et al. Suicide attempters with Borderline Personality Disorder show differential orbitofrontal and parietal recruitment when reflecting on aversive memories. J Psychiatr Res 2016; : 71–78. 51 Zhang R, Geng X, Lee TMC. Large-scale functional neural network correlates of response inhibition: an fMRI meta-analysis. Brain Struct Funct 2017; 222: 3973–3990. 52 Minzenberg MJ, Lesh TA, Niendam TA, Yoon JH, Cheng Y, Rhoades RN et al. Control-related frontal-striatal function is associated with past suicidal ideation and behavior in patients with recent-onset psychotic major mood disorders. J Affect Disord 2015; 188: 202–209. 53 Matthews S, Spadoni A, Knox K, Strigo I, Simmons A. Combat-exposed war veterans at risk for suicide show hyperactivation of prefrontal cortex and anterior cingulate during error processing. Psychosom Med 2012; 74: 471–475. 54 Minzenberg MJ, Lesh TA, Niendam TA, Yoon JH, Rhoades RN, Carter CS. Frontal cortex control dysfunction related to long-term suicide risk in recent-onset schizophrenia. Schizophr Res 2014; 157: 19–25. 55 Fradkin Y, Khadka S, Bessette KL, Stevens MC. The relationship of impulsivity and cortical thickness in depressed and non-depressed adolescents. Brain Imaging Behav 2017; 11: 1515–1525. 56 Winkler AM, Kochunov P, Blangero J, Almasy L, Zilles K, Fox PT et al. Cortical thickness or grey matter volume? The importance of selecting the phenotype for imaging genetics studies. Neuroimage 2010; 53: 1135–1146. 57 Rakic P. A small step for the cell, a giant leap for mankind: a hypothesis of neocortical expansion during evolution. Trends Neurosci 1995; 18: 383–388. 6 58 Wagner G, Schultz CC, Koch K, Schachtzabel C, Sauer H, Schlösser RG. Prefrontal cortical thickness in depressed patients with high-risk for suicidal behavior. J Psychiatr Res 2012; 46: 1449–1455. 59 Rakic P. Specification of Cerebral Cortical Areas. Science 1988; 241: 170–176. 60 Just MA, Pan L, Cherkassky VL, Mcmakin D, Cha C, Nock MK et al. Machine learning of neural representations of suicide and emotion concepts identifies suicidal youth. Nat Hum Behav 2017; 1: 911–919. 61 Fellows LK, Farah MJ. Different underlying impairments in decision-making following ventromedial and dorsolateral frontal lobe damage in humans. Cereb Cortex 2005; 15: 58–63. 62 Richard-Devantoy S, Berlim MT, Jollant F. A meta-analysis of neuropsychological markers of vulnerability to suicidal behavior in mood disorders. Psychol Med 2014; 44: 1663–1673. 63 Andrews-Hanna JR, Smallwood J, Spreng RN. The default network and self- generated thought: Component processes, dynamic control, and clinical relevance. Ann N Y Acad Sci 2014; 1316: 29–52. 64 Wagner DD, Haxby J V., Heatherton TF. The representation of self and person knowledge in the medial prefrontal cortex. Wiley Interdiscip Rev Cogn Sci 2012; 3: 451–470. 65 Miller EK, Cohen JD. An integrative theory of prefrontal cortex function. Annu Rev Neurosci 2001; 24: 167–202. 66 Rive MM, Van Rooijen G, Veltman DJ, Mary ML, Schene AH, Ruhé HG. Neural correlates of dysfunctional emotion regulation in major depressive disorder. A systematic review of neuroimaging studies. Neurosci Biobehav Rev 2013; 37: 2529– 2553. 67 Petrides M. Lateral prefrontal cortex: Architectonic and functional organization. Philos Trans R Soc B Biol Sci 2005; 360: 781–795. 68 Baird B, Smallwood J, Gorgolewski KJ, Margulies DS. Medial and Lateral Networks in Anterior Prefrontal Cortex Support Metacognitive Ability for Memory and Perception. J Neurosci 2013; 33: 16657–16665. 69 De Martino B, Fleming SM, Garrett N, Dolan RJ. Confidence in value-based choice. Nat Neurosci 2013; 16: 105–110. 70 Hwang JP, Lee TW, Tsai SJ, Chen TJ, Yang CH, Lirng JF et al. Cortical and subcortical abnormalities in late-onset depression with history of suicide attempts investigated with MRI and voxel-based morphometry. J Geriatr Psychiatry Neurol 2010; 23: 171–184. 71 Besteher B, Wagner G, Koch K, Schachtzabel C, Reichenbach JR, Schlösser R et al. Pronounced prefronto-temporal cortical thinning in schizophrenia: Neuroanatomical correlate of suicidal behavior? Schizophr Res 2016; 176: 151–157. 72 Reisch T, Seifritz E, Esposito F, Wiest R, Valach L, Michel K. An fMRI study on mental 7 pain and suicidal behavior. J Affect Disord 2010; 126: 321–325. 73 Miller AB, McLaughlin KA, Busso DS, Brueck S, Peverill M, Sheridan MA. Neural Correlates of Emotion Regulation and Adolescent Suicidal Ideation. Biol Psychiatry Cogn Neurosci Neuroimaging 2018; 3: 125–132. 74 Pan LA, Hassel S, Segreti AM, Nau SA, Brent DA, Phillips ML. Differential patterns of activity and functional connectivity in emotion processing neural circuitry to angry and happy faces in adolescents with and without suicide attempt. Psychol Med 2013; 43: 2129–2142. 75 Hagan CR, Joiner TE. The Indirect Effect of Perceived Criticism on Suicide Ideation and Attempts. Arch Suicide Res 2017; 21: 438–454. 76 Hames JL, Rogers ML, Silva C, Ribeiro JD, Teale NE, Joiner TE. A Social Exclusion Manipulation Interacts with Acquired Capability for Suicide to Predict Self-Aggressive Behaviors. Arch Suicide Res 2018; 22: 32–45. 77 Tsutsui KI, Grabenhorst F, Kobayashi S, Schultz W. A dynamic code for economic object valuation in prefrontal cortex neurons. Nat Commun 2016; 7. doi:10.1038/ncomms12554. 78 Vanyukov PM, Szanto K, Hallquist MN, Siegle GJ, Reynolds CF, Forman SD et al. Paralimbic and lateral prefrontal encoding of reward value during intertemporal choice in attempted suicide. Psychol Med 2016; 46: 381–391. 79 Dombrovski AY, Szanto K, Siegle GJ, Wallace ML, Forman SD, Sahakian B et al. Lethal forethought: Delayed reward discounting differentiates high- and low-lethality suicide attempts in old age. Biol Psychiatry 2011; 70: 138–144. 80 Minzenberg MJ, Lesh T, Niendam T, Yoon JH, Cheng Y, Rhoades RN et al. Frontal motor cortex activity during reactive control is associated with past suicidal behavior in recent-onset schizophrenia. Crisis 2015; 36: 363–370. 81 Lee KH, Pluck G, Lekka N, Horton A, Wilkinson ID, Woodruff PWR. Self-harm in schizophrenia is associated with dorsolateral prefrontal and posterior cingulate activity. Prog Neuro-Psychopharmacology Biol Psychiatry 2015; 61: 18–23. 82 Willeumier K, Taylor D V., Amen DG. Decreased cerebral blood flow in the limbic and prefrontal cortex using SPECT imaging in a cohort of completed suicides. Transl Psychiatry 2011; 1: e28-8. 83 Sublette ME, Milak MS, Galfalvy HC, Oquendo MA, Malone KM, Mann JJ. Regional Brain Glucose Uptake Distinguishes Suicide Attempters from Non-Attempters in Major Depression. Arch Suicide Res 2013; 17: 434–447. 84 Oquendo MA, Placidi GPA, Malone KM, Campbell C, Keilp J, Brodsky B et al. Positron emission tomography of regional brain metabolic responses to a serotonergic challenge and lethality of suicide attempts in major depression. Arch Gen Psychiatry 2003; 60: 14–22. 85 van Heeringen K, Wu GR, Vervaet M, Vanderhasselt MA, Baeken C. Decreased resting state metabolic activity in frontopolar and parietal brain regions is associated 8 with suicide plans in depressed individuals. J Psychiatr Res 2017; 84: 243–248. 86 Oquendo MA. Impulsive versus planned suicide attempts: Different phenotypes? J Clin Psychiatry 2015; 76: 293–294. 87 Goodman M, Hazlett EA, Avedon JB, Siever DR, Chu KW, New AS. Anterior cingulate volume reduction in adolescents with borderline personality disorder and co-morbid major depression. J Psychiatr Res 2011; 45: 803–807. 88 Soloff P, White R, Diwadkar VA. Impulsivity, aggression and brain structure in high and low lethality suicide attempters with borderline personality disorder. Psychiatry Res - Neuroimaging 2014; 222: 131–139. 89 Kolla NJ, Chiuccariello L, Wilson AA, Houle S, Links P, Bagby RM et al. Elevated Monoamine Oxidase-A Distribution Volume in Borderline Personality Disorder Is Associated with Severity Across Mood Symptoms, Suicidality, and Cognition. Biol Psychiatry 2016; 79: 117–126. 90 Holmes SE, Hinz R, Conen S, Gregory CJ, Matthews JC, Anton-Rodriguez JM et al. Elevated Translocator Protein in Anterior Cingulate in Major Depression and a Role for Inflammation in Suicidal Thinking: A Positron Emission Tomography Study. Biol Psychiatry 2018; 83: 61–69. 91 Prescot A, Sheth C, Legarreta M, Renshaw PF, McGlade E, Yurgelun-Todd D. Altered Cortical Gamma-Amino Butyric Acid in Female Veterans With Suicidal Behavior: Sex Differences and Clinical Correlates. Chronic Stress 2018; 2. doi:10.1177/2470547018768771. 92 Baek K, Kwon J, Chae JH, Chung YA, Kralik JD, Min JA et al. Heightened aversion to risk and loss in depressed patients with a suicide attempt history. Sci Rep 2017; 7: 11228. 93 Dombrovski AY, Szanto K, Clark L, Reynolds CF, Siegle GJ. Reward signals, attempted suicide, and impulsivity in late-life depression. JAMA Psychiatry 2013; 70: 1020–1030. 94 Uddin LQ, Kinnison J, Pessoa K, Anderson ML. Beyond the Tripartite Cognition– Emotion–Interoception Model of the Human Insular Cortex. J Cogn Neurosci 2014; 26: 16–27. 95 Craig AD. Interoception: The sense of the physiological condition of the body. Curr Opin Neurobiol 2003; 13: 500–505. 96 Singer T, Seymour B, Doherty JPO, Stephan KE, Dolan RJ, Frith CD. Understanding the Emergence of Neuropsychiatric Disorders with Network[1]. 2009; 439: 466–469. 97 Singer T, Seymour B, O’Doherty J, Kaube H, Dolan RJ, Frith CD. Empathy for Pain Involves the Affective but not Sensory Components of Pain. Science (80- ) 2004; 303: 1157–1162. 98 Forrest LN, Smith AR, White RD, Joiner TE. (Dis)connected: An examination of interoception in individuals with suicidality. J Abnorm Psychol 2015; 124: 754–763. 99 Brausch AM, Woods SE. Emotion Regulation Deficits and Nonsuicidal Self-Injury 9 Prospectively Predict Suicide Ideation in Adolescents. Suicide Life-Threatening Behav 2018. doi:10.1111/sltb.12478. 100 Soloff PH, Pruitt P, Sharma M, Radwan J, White R, Diwadkar VA. Structural brain abnormalities and suicidal behavior in borderline personality disorder. J Psychiatr Res 2012; 46: 516–525. 101 Duarte DGG, Neves M de CL, Albuquerque MR, Turecki G, Ding Y, de Souza-Duran FL et al. Structural brain abnormalities in patients with type I bipolar disorder and suicidal behavior. Psychiatry Res - Neuroimaging 2017; 265: 9–17. 102 Amen DG, Prunella JR, Fallon JH, Amen B, Hanks C. A Comparative Analysis of Completed Suicide Using High Resolution Brain SPECT Imaging. J Neuropsychiatry Clin Neurosci 2009; 21: 430–439. 103 Olié E, Jollant F, Deverdun J, De Champfleur NM, Cyprien F, Le Bars E et al. The experience of social exclusion in women with a history of suicidal acts: A neuroimaging study. Sci Rep 2017; 7: 1–8. 104 Savitz J, Harrison NA. Interoception and Inflammation in Psychiatric Disorders. Biol Psychiatry Cogn Neurosci Neuroimaging 2018; 3: 514–524. 105 Canteras NS, Swanson LW. Projections of the ventral subiculum to the amygdala, septum, and hypothalamus: A PHAL anterograde tract‐tracing study in the rat. J Comp Neurol 1992; 324: 180–194. 106 Zola-Morgan. Effects of lesions of perirhinal cortex or parahippocampal cortex on memory in monkeys. Soc Neurosci Abstr 1994; 9: 4355–4370. 107 Patel D, Anilkumar S, Chattarji S, Buwalda B. Repeated social stress leads to contrasting patterns of structural plasticity in the amygdala and hippocampus. Behav Brain Res 2018; 347: 314–324. 108 Doré BP, Rodrik O, Boccagno C, Hubbard A, Weber J, Stanley B et al. Negative Autobiographical Memory in Depression Reflects Elevated Amygdala-Hippocampal Reactivity and Hippocampally Associated Emotion Regulation. Biol Psychiatry Cogn Neurosci Neuroimaging 2018; 3: 358–366. 109 Monkul ES, Hatch JP, Nicoletti MA, Spence S, Brambilla P, Lacerda ALT et al. Fronto- limbic brain structures in suicidal and non-suicidal female patients with major depressive disorder. Mol Psychiatry 2007; 12: 360–366. 110 Spoletini I, Piras F, Fagioli S, Rubino IA, Martinotti G, Siracusano A et al. Suicidal attempts and increased right amygdala volume in schizophrenia. Schizophr Res 2011; 125: 30–40. 111 Rentería ME, Schmaal L, Hibar DP, Couvy-Duchesne B, Strike LT, Mills NT et al. Subcortical brain structure and suicidal behaviour in major depressive disorder: A meta-analysis from the ENIGMA-MDD working group. Transl Psychiatry 2017; 7: e1116. 112 Gifuni AJ, Ding Y, Olié E, Lawrence N, Cyprien F, Le Bars E et al. Subcortical nuclei volumes in suicidal behavior: nucleus accumbens may modulate the lethality of acts. 10 Brain Imaging Behav 2016; 10: 96–104. 113 Lijffijt M, Rourke ED, Swann AC, Zunta-Soares GB, Soares JC. Illness-course modulates suicidality-related prefrontal gray matter reduction in women with bipolar disorder. Acta Psychiatr Scand 2014; 130: 374–387. 114 Colle R, Chupin M, Cury C, Vandendrie C, Gressier F, Hardy P et al. Depressed suicide attempters have smaller hippocampus than depressed patients without suicide attempts. J Psychiatr Res 2015; 61: 13–18. 115 Pan LA, Ramos L, Segreti AM, Brent DA, Phillips ML. Right superior temporal gyrus volume in adolescents with a history of suicide attempt. Br J Psychiatry 2015; 206: 339–340. 116 Soloff PH, Price JC, Meltzer CC, Fabio A, Frank GK, Kaye WH. 5HT2A Receptor Binding is Increased in Borderline Personality Disorder. Biol Psychiatry 2007; 62: 580– 587. 117 Marchand WR, Lee JN, Garn C, Thatcher J, Gale P, Kreitschitz S et al. Aberrant emotional processing in posterior cortical midline structures in bipolar II depression. Prog Neuro-Psychopharmacology Biol Psychiatry 2011; 35: 1729–1737. 118 Spreng RN, Mar RA, Kim ASN. The common neural basis of autobiographical memory, prospection, navigation, theory of mind, and the default mode: A quantitative meta-analysis. J Cogn Neurosci 2009; 21: 489–510. 119 Schacter DL, Addis DR, Buckner RL. Remembering the past to imagine the future: the prospective brain. Nat Rev Neurosci 2007; 8: 657. 120 Peters J, Büchel C. The neural mechanisms of inter-temporal decision-making: Understanding variability. Trends Cogn Sci 2011; 15: 227–239. 121 Peters J, Büchel C. Episodic Future Thinking Reduces Reward Delay Discounting through an Enhancement of Prefrontal-Mediotemporal Interactions. Neuron 2010; 66: 138–148. 122 van Erp TGM, Hibar DP, Rasmussen JM, Glahn DC, Pearlson GD, Andreassen O a et al. Subcortical brain volume abnormalities in 2028 individuals with schizophrenia and 2540 healthy controls via the ENIGMA consortium. Mol Psychiatry 2015; : 1–7. 123 Schmaal L, Veltman DJ, Van Erp TGM, Smann PG, Frodl T, Jahanshad N et al. Subcortical brain alterations in major depressive disorder: Findings from the ENIGMA Major Depressive Disorder working group. Mol Psychiatry 2016; 21: 806–812. 124 Hibar DP, Westlye LT, Van Erp TGM, Rasmussen J, Leonardo CD, Faskowitz J et al. Subcortical volumetric abnormalities in bipolar disorder. Mol Psychiatry 2016; 21: 1710–1716. 125 Cummings JL. Frontal-Subcortical Circuits and Human Behavior. Arch Neurol 1993; 50: 873–880. 126 Bracht T, Linden D, Keedwell P. A review of white matter microstructure alterations of pathways of the reward circuit in depression. J Affect Disord 2015; 187: 45–53. 127 Featherstone RE, McDonald RJ. Dorsal striatum and stimulus-response learning: 11 Lesions of the dorsolateral, but not dorsomedial, striatum impair acquisition of a stimulus-response-based instrumental discrimination task, while sparing conditioned place preference learning. Neuroscience 2004; 124: 23–31. 128 Alitto, H. J., & Usrey WM (2003). Corticothalamic feedback and sensory processing. Curr Opin Neurobiol 2003; 13: 440–445. 129 Dombrovski AY, Siegle GJ, Szanto K, Clark L, Reynolds CF, Aizenstein H. The temptation of suicide: Striatal gray matter, discounting of delayed rewards, and suicide attempts in late-life depression. Psychol Med 2012; 42: 1203–1215. 130 Vang FJ, Ryding E, Träskman-Bendz L, van Westen D, Lindström MB. Size of basal ganglia in suicide attempters, and its association with temperament and serotonin transporter density. Psychiatry Res - Neuroimaging 2010; 183: 177–179. 131 Nye JA, Purselle D, Plisson C, Voll RJ, Stehouwer JS, Votaw JR et al. Decreased brainstem and putamen sert binding potential in depressed suicide attempters using [11C]-zient pet imaging. Depress Anxiety 2013; 30: 902–907. 132 Ryding E, Ahnlide JA, Lindström M, Rosén I, Träskman-Bendz L. Regional brain serotonin and dopamine transporter binding capacity in suicide attempters relate to impulsiveness and mental energy. Psychiatry Res - Neuroimaging 2006; 148: 195– 203. 133 Yeh YW, Ho PS, Chen CY, Kuo SC, Liang CS, Yen CH et al. Suicidal ideation modulates the reduction in serotonin transporter availability in male military conscripts with major depression: A 4-[18F]-ADAM PET study. World J Biol Psychiatry 2015; 16: 502–512. 134 Lopez-Larson M, King JB, McGlade E, Bueler E, Stoeckel A, Epstein DJ et al. Enlarged thalamic volumes and increased fractional anisotropy in the thalamic radiations in veterans with suicide behaviors. Front Psychiatry 2013; 4: 1–13. 135 Marchand WR, Lee JN, Garn C, Thatcher J, Gale P, Kreitschitz S et al. Striatal and cortical midline activation and connectivity associated with suicidal ideation and depression in bipolar II disorder. J Affect Disord 2011; 133: 638–645. 136 Kim YJ, Park HJ, Jahng GH, Lee SM, Kang WS, Kim SK et al. A pilot study of differential brain activation to suicidal means and DNA methylation of CACNA1C gene in suicidal attempt patients. Psychiatry Res 2017; 255: 42–48. 137 Richard-Devantoy S, Olié E, Guillaume S, Courtet P. Decision-making in unipolar or bipolar suicide attempters. J Affect Disord 2016; 190: 128–136. 138 Marchand WR, Lee JN, Johnson S, Thatcher J, Gale P, Wood N et al. Striatal and cortical midline circuits in major depression: Implications for suicide and symptom expression. Prog Neuro-Psychopharmacology Biol Psychiatry 2012; 36: 290–299. 139 Booth MCA, Rolls ET. View-invariant representations of familiar objects by neurons in the inferior temporal visual cortex. Cereb Cortex 1998; 8: 510–523. 140 Haxby J V, Hoffman EA, Gobbini MI. The distributed human neural system for face perception. Trends Cogn Sci 2000; 4. 12 141 Howard M a, Volkov IO, Mirsky R, Garell PC, Noh MD, Granner M et al. Auditory cortex on the human posterior superior temporal gyrus. J Comp Neurol 2000; 416: 79– 92. 142 Aguilar EJ, García-Martí G, Martí-Bonmatí L, Lull JJ, Moratal D, Escartí MJ et al. Left orbitofrontal and superior temporal gyrus structural changes associated to suicidal behavior in patients with schizophrenia. Prog Neuro-Psychopharmacology Biol Psychiatry 2008; 32: 1673–1676. 143 Audenaert K, Goethals I, Van laere K, Lahorte P, Brans B, Versijpt J et al. Spect neuropsychological activation procedure with the verbal fluency test in attempted suicide patients. Nucl Med Commun 2002; 23: 907–916. 144 Motoyama H, Hishitani S. The brain mechanism that reduces the vividness of negative imagery. Conscious Cogn 2016; 39: 59–69. 145 Feng C, Yan X, Huang W, Han S, Ma Y. Neural representations of the multidimensional self in the cortical midline structures. Neuroimage 2018; 183: 291– 299. 146 Peng H, Wu K, Li J, Qi H, Guo S, Chi M et al. Increased suicide attempts in young depressed patients with abnormal temporal-parietal-limbic gray matter volume. J Affect Disord 2014; 165: 69–73. 147 Quevedo K, Ng R, Scott H, Martin J, Smyda G, Keener M et al. The Neurobiology of Self-Face Recognition in Depressed Adolescents with Low or High Suicidality. J Abnorm Psychol 2016; 125: 1185–1200. 148 Adamaszek M, D’Agata F, Ferrucci R, Habas C, Keulen S, Kirkby KC et al. Consensus Paper: Cerebellum and Emotion. Cerebellum 2017; 16: 552–576. 149 Womer FY, Wang F, Chepenik LG, Kalmar JH, Spencer L, Edmiston E et al. Sexually dimorphic features of vermis morphology in bipolar disorder. Bipolar Disord 2009; 11: 753–758. 150 Lee YJ, Kim S, Gwak AR, Kim SJ, Kang SG, Na KS et al. Decreased regional gray matter volume in suicide attempters compared to suicide non-attempters with major depressive disorders. Compr Psychiatry 2016; 67: 59–65. 151 Ballard ED, Lally N, Nugent AC, Furey ML, Luckenbaugh DA, Zarate CA. Neural correlates of suicidal ideation and its reduction in depression. Int J Neuropsychopharmacol 2015; 18: 1–6. 152 Ordaz SJ, Goyer MS, Ho TC, Singh MK, Gotlib IH. Network basis of suicidal ideation in depressed adolescents. J Affect Disord 2018; 226: 92–99. 153 Du L, Zeng J, Liu H, Tang D, Meng H, Li Y et al. Fronto-limbic disconnection in depressed patients with suicidal ideation: A resting-state functional connectivity study. J Affect Disord 2017; 215: 213–217. 154 Zhang S, Chen J mei, Kuang L, Cao J, Zhang H, Ai M et al. Association between abnormal default mode network activity and suicidality in depressed adolescents. BMC Psychiatry 2016; 16: 1–10. 13 155 Beaulieu C. The basis of anisotropic water diffusion in the nervous system - A technical review. NMR Biomed 2002; 15: 435–455. 156 Mahon K, Burdick KE, Wu J, Ardekani BA, Szeszko PR. Relationship between suicidality and impulsivity in bipolar I disorder: A diffusion tensor imaging study. Bipolar Disord 2012; 14: 80–89. 157 Cyprien F, de Champfleur NM, Deverdun J, Olié E, Le Bars E, Bonafé A et al. Corpus callosum integrity is affected by mood disorders and also by the suicide attempt history: A diffusion tensor imaging study. J Affect Disord 2016; 206: 115–124. 158 Lischke A, Domin M, Freyberger HJ, Grabe HJ, Mentel R, Bernheim D et al. Structural Alterations in the Corpus Callosum Are Associated with Suicidal Behavior in Women with Borderline Personality Disorder. Front Hum Neurosci 2017; 11: 1–10. 159 Laird, P. Mickle Fox, Simon B. Eickhoff, Jessica A. Turner, Kimberly L. Ray, D. Reese McKay, David C. Glahn, Christian F. Beckmann, Stephen M. Smith and PTF. Behavioral Interpretations of Intrinsic Connectivity Networks. J Cogn Neurosci 2011; 23: 4022–4037. 160 Kang SG, Na KS, Choi JW, Kim JH, Son YD, Lee YJ. Resting-state functional connectivity of the amygdala in suicide attempters with major depressive disorder. Prog Neuro-Psychopharmacology Biol Psychiatry 2017; 77: 222–227. 161 Quevedo K, Ng R, Scott H, Kodavaganti S, Smyda G, Diwadkar V et al. Ventral Striatum Functional Connectivity during Rewards and Losses and Symptomatology in Depressed Patients. Biol Psychol 2017; 123: 62–73. 162 Kim K, Kim SW, Myung W, Han CE, Fava M, Mischoulon D et al. Reduced orbitofrontal-thalamic functional connectivity related to suicidal ideation in patients with major depressive disorder. Sci Rep 2017; 7: 15772. 163 Myung W, Han CE, Fava M, Mischoulon D, Papakostas GI, Heo JY et al. Reduced frontal-subcortical white matter connectivity in association with suicidal ideation in major depressive disorder. Transl Psychiatry 2016; 6: e835--8. 164 Bijttebier S, Caeyenberghs K, van den Ameele H, Achten E, Rujescu D, Titeca K et al. The Vulnerability to Suicidal Behavior is Associated with Reduced Connectivity Strength. Front Hum Neurosci 2015; 9. doi:10.3389/fnhum.2015.00632. 165 Jia Z, Huang X-Q, Wu Q-Z, Zhang T-J, Lui S, Zhang J et al. High-field magnetic resonance imaging of suicidality in patients with major depressive disorder. Am J Psychiatry 2010; 167: 1381–1390. 166 Jia Z, Wang Y, Huang X, Kuang W, Wu Q, Lui S et al. Impaired frontothalamic circuitry in suicidal patients with depression revealed by diffusion tensor imaging at 3.0 T. J Psychiatry Neurosci 2014; 39: 170–177. 167 Bressler SL, Menon V. Large-scale brain networks in cognition: emerging methods and principles. Trends Cogn Sci 2010; 14: 277–290. 168 Cao J, Chen X, Chen J, Ai M, Gan Y, Wang W et al. Resting-state functional MRI of abnormal baseline brain activity in young depressed patients with and without suicidal 14 behavior. J Affect Disord 2016; 205: 252–263. 169 Olvet DM, Peruzzo D, Thapa-Chhetry B, Sublette ME, Sullivan GM, Oquendo MA et al. A diffusion tensor imaging study of suicide attempters. J Psychiatr Res 2014; 51: 60–67. 170 Minzenberg M, Lesh T, Niendam T, Yoon J, Cheng Y, Rhoades R et al. Conflict- related anterior cingulate functional connectivity is associated with past suicidal ideation and behavior in recent-onset schizophrenia. J Psychiatr Res 2015; 65: 95– 101. 171 Minzenberg MJ, Lesh T, Niendam T, Yoon JH, Cheng Y, Rhoades R et al. Conflict- related anterior cingulate functional connectivity is associated with past suicidal ideation and behavior in recent-onset Psychotic Major Mood Disorders. J Neuropsychiatry Clin Neurosci 2016; 28: 95–101. 172 Janes AC, Farmer S, Peechatka AL, Frederick BDB, Lukas SE. Insula-dorsal anterior cingulate cortex coupling is associated with enhanced brain reactivity to smoking cues. Neuropsychopharmacology 2015; 40: 1561–1568. 173 Giesecke T, Gracely RH, Williams DA, Geisser ME, Petzke FW, Clauw DJ. The relationship between depression, clinical pain, and experimental pain in a chronic pain cohort. Arthritis Rheum 2005; 52: 1577–1584. 174 Stevens F, Hurley R, Taber K. Anterior Cingulate Cortex: Unique Role in Cognition and Emotion. J neuropsychiatry Clin Neurosci 2011; 23: 121–125. 175 Gasquoine PG. Contributions of the insula to cognition and emotion. Neuropsychol Rev 2014; 24: 77–87. 176 Menon V, Uddin LQ. Saliency, switching, attention and control: a network model of insula function. Brain Struct Funct 2010; 214: 655–667. 177 Nieuwenhuys R. The insular cortex. A review. 1st ed. Elsevier B.V., 2012 doi:10.1016/B978-0-444-53860-4.00007-6. 178 Deshpande G, Baxi M, Witte T, Robinson JL. A neural basis for the acquired capability for suicide. Front Psychiatry 2016; 7: 1–19. 179 Goulden N, Khusnulina A, Davis NJ, Bracewell RM, Bokde AL, McNulty JP et al. The salience network is responsible for switching between the default mode network and the central executive network: Replication from DCM. Neuroimage 2014; 99: 180–190. 180 Zhou Y, Friston KJ, Zeidman P, Chen J, Li S, Razi A. The Hierarchical Organization of the Default, Dorsal Attention and Salience Networks in Adolescents and Young Adults. Cereb Cortex 2018; 28: 726–737. 181 Sridharan D, Levitin D, Menon V. A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks. Proc Natl Acad Sci 2008; 105: 12569–12574. 182 Cox Lippard ET, Johnston JAY, Spencer L, Quatrano S, Fan S, Sankar A et al. Preliminary examination of gray and white matter structure and longitudinal structural changes in frontal systems associated with future suicide attempts in adolescents and 15 young adults with mood disorders. J Affect Disord 2018; in press. 183 Wilkinson ST, Sanacora G. Ketamine: a Potential Rapid-Acting Antisuicidal Agent? Depress Anxiety 2016; 33: 711–717. 184 Cipriani A, Hawton K, Stockton S, Geddes JR. Lithium in the prevention of suicide in mood disorders: updated systematic review and meta-analysis. BMJ 2013; 346: f3646. 185 Al Jurdi RK, Swann A, Mathew SJ. Psychopharmacological Agents and Suicide Risk Reduction: Ketamine and Other Approaches. Curr Psychiatry Rep 2015; 17. doi:10.1007/s11920-015-0614-9. 186 Kundu P, Benson BE, Rosen D, Frangou S, Leibenluft E, Luh W-M et al. The integration of functional brain activity from adolescence to adulthood. J Neurosci 2018; 38: 1864–17. 187 Kochunov P, Glahn DC, Lancaster J, Thompson PM, Kochunov V, Rogers B et al. Fractional anisotropy of cerebral white matter and thickness of cortical gray matter across the lifespan. Neuroimage 2011; 58: 41–49. 188 Fan S, Lippard E, Sankar A, Wallace A, Johnston J, Wang F et al. Grey and white matter differences in adolescents and young adults with prior suicide attempts across bipolar and major depressive disorders. J Affect Disord 2018; in press. 189 Fox KR, Millner AJ, Mukerji CE, Nock MK. Examining the role of sex in self-injurious thoughts and behaviors. Clin Psychol Rev 2017; : 0–1. 190 Olfson M, Blanco C, Wall M, Liu SM, Saha TD, Pickering RP et al. National Trends in Suicide Attempts Among Adults in the United States. JAMA psychiatry 2017; 74: 1095–1103. 191 Choi NG, DiNitto DM, Marti CN, Kaplan MS, Conwell Y. Suicide Means among Decedents Aged 50+ Years, 2005–2014: Trends and Associations with Sociodemographic and Precipitating Factors. Am J Geriatr Psychiatry 2017; 25: 1404– 1414. 192 Kaplan MS, McFarland BH, Huguet N. Characteristics of adult male and female firearm suicide decedents: Findings from the National Violent Death Reporting System. Inj Prev 2009; 15: 322–327. 193 Hedegaard H, Curtin S, Warner M. Suicide Mortality in the United States, 1999-2017. NCHS Data Brief 2018; 330: 1–8. 194 Nery-Fernandes F, Rocha M V., Jackowski A, Ladeia G, Guimarães JL, Quarantini LC et al. Reduced posterior corpus callosum area in suicidal and non-suicidal patients with bipolar disorder. J Affect Disord 2012; 142: 150–155. 195 Gabbay V, Bradley KA, Mao X, Ostrover R, Kang G, Shungu DC. Anterior cingulate cortex γ-aminobutyric acid deficits in youth with depression. Transl Psychiatry 2017; 7. doi:10.1038/tp.2017.187. 196 Soloff PH, Meltzer CC, Becker C, Greer PJ, Kelly TM, Constantine D. Impulsivity and prefrontal hypometabolism in borderline personality disorder. Psychiatry Res 2003; 123: 153–163. 16 197 Soloff PH, Chiappetta L, Mason NS, Becker C, Price JC. Effects of serotonin-2A receptor binding and gender on personality traits and suicidal behavior in borderline personality disorder. Psychiatry Res 2014; 222: 140–148. 198 Ehrlich S, Noam GG, Lyoo IK, Kwon BJ, Clark MA, Renshaw PF. White matter hyperintensities and their associations with suicidality in psychiatrically hospitalized children and adolescents. J Am Acad Child Adolesc Psychiatry 2004; 43: 770–776. 199 Chase HW, Segreti AM, Keller TA, Cherkassky VL, Just MA, Pan LA et al. Alterations of functional connectivity and intrinsic activity within the cingulate cortex of suicidal ideators. J Affect Disord 2017; 212: 78–85. 200 di Giacomo E, Krausz M, Colmegna F, Aspesi F, Clerici M. Estimating the risk of attempted suicide among sexual minority youths: a systematic review and meta- analysis. JAMA Pediatr 2018; 172: 1145–1152. 201 Mueller S, De Cuypere G, T’Sjoen G. Trnasgender research in the 21st century: a selective critical review from a neurocognitive perspective. Am J Psychiatry 2017; 174: 1155–1162. 202 Woo CW, Chang LJ, Lindquist MA, Wager TD. Building better biomarkers: Brain models in translational neuroimaging. Nat Neurosci 2017; 20: 365–377. 203 Duyn JH. The future of ultra-high field MRI and fMRI for study of the human brain. Neuroimage 2012; 62: 1241–1248. 204 Courtet P, Giner L, Seneque M, Guillaume S, Olie E, Ducasse D. Neuroinflammation in suicide: Toward a comprehensive model. World J Biol Psychiatry 2016; 17: 564– 586. 205 Oquendo MA, Sullivan GM, Sudol K, Baca-Garcia E, Stanley BH, Sublette ME et al. Toward a biosignature for suicide. Am J Psychiatry 2014; 171: 1259–1277. 206 Sequeira A, Mamdani F, Ernst C, Vawter MP, Bunney WE, Lebel V et al. Global brain gene expression analysis links Glutamatergic and GABAergic alterations to suicide and major depression. PLoS One 2009; 4: 21–23. 207 Zhao J, Verwer RWH, Gao SF, Qi XR, Lucassen PJ, Kessels HW et al. Prefrontal alterations in GABAergic and glutamatergic gene expression in relation to depression and suicide. J Psychiatr Res 2018; 102: 261–274. 208 Poulter MO. Altered organization of GABAA receptor mRNA expression in the depressed suicide brain. Front Mol Neurosci 2010; 3: 1–10. 209 Lebowitz E, Blumberg H, Silverman W. Negative peer social interactions and oxytocin levels linked to suicidal ideation in anxious youth. J Affect Disord 2018; in press. 210 Miranda R, Nolen-Hoeksema S. Brooding and reflection: Rumination predicts suicidal ideation at 1-year follow-up in a community sample. Behav Res Ther 2007; 45: 3088– 3095. 211 Williams JMG, Van Der Does AJW, Barnhofer T, Crane C, Segal ZS. Cognitive reactivity, suicidal ideation and future fluency: Preliminary investigation of a differential activation theory of hopelessness/suicidality. Cognit Ther Res 2008; 32: 83–104. 17 212 Auerbach RP, Millner AJ, Stewart JG, Esposito EC. Identifying differences between depressed adolescent suicide ideators and attempters. J Affect Disord 2015; 186: 127–133. 213 Peyre H, Hoertel N, Stordeur C, Lebeau G, Blanco C, McMahon K et al. Contributing Factors and Mental Health Outcomes of First Suicide Attempt During Childhood and Adolescence. J Clin Psychiatry 2015; 78: e622. 214 Björkenstam C, Kosidou K, Björkenstam E. Childhood adversity and risk of suicide: cohort study of 548 721 adolescents and young adults in Sweden. BMJ 2017; 357: j1334. 215 van Harmelen A-L, van Tol M-J, van der Wee NJA, Veltman DJ, Aleman A, Spinhoven P et al. Reduced Medial Prefrontal Cortex Volume in Adults Reporting Childhood Emotional Maltreatment. Biol Psychiatry 2010; 68: 832–838. 216 Van Harmelen AL, Hauber K, Moor BG, Spinhoven P, Boon AE, Crone EA et al. Childhood emotional maltreatment severity is associated with dorsal medial prefrontal cortex responsivity to social exclusion in young adults. PLoS One 2014; 9: e85107. 217 Button K, Ioannidis J, Mokrysz C, Nosek B, Flint J, Robinson E et al. Power failure: why small sample size undermines the reliability of neuroscience. Nat Rev Neurosci 2013; 14: 365–375. 218 Fischer HF, Rose M. Scoring Depression on a Common Metric: A Comparison of EAP Estimation, Plausible Value Imputation, and Full Bayesian IRT Modeling. Multivariate Behav Res 2018; 54: 85–99. 219 Rüsch N, Spoletini I, Wilke M, Martinotti G, Bria P, Trequattrini A et al. Inferior frontal white matter volume and suicidality in schizophrenia. Psychiatry Res - Neuroimaging 2008; 164: 206–214. 220 Matsuo K, Nielsen N, Nicoletti MA, Hatch JP, Monkul ES, Watanabe Y et al. Anterior genu corpus callosum and impulsivity in suicidal patients with bipolar disorder. Neurosci Lett 2010; 469: 75–80. 221 Gifuni AJ, Olié E, Ding Y, Cyprien F, le Bars E, Bonafé A et al. Corpus callosum volumes in bipolar disorders and suicidal vulnerability. Psychiatry Res - Neuroimaging 2017; 262: 47–54. 222 Harenski CL, Brook M, Kosson DS, Bustillo JR, Harenski KA, Caldwell MF et al. Socio-neuro risk factors for suicidal behavior in criminal offenders with psychotic disorders. Soc Cogn Affect Neurosci 2017; 12: 70–80. 223 Baldaçara L, Nery-Fernandes F, Rocha M, Quarantini LC, Rocha GGL, Guimarães JL et al. Is cerebellar volume related to bipolar disorder? J Affect Disord 2011; 135: 305– 309. 224 Benedetti F, Riccaboni R, Poletti S, Radaelli D, Locatelli C, Lorenzi C et al. The serotonin transporter genotype modulates the relationship between early stress and adult suicidality in bipolar disorder. Bipolar Disord 2014; 16: 857–866. 225 Cyprien F, Courtet P, Malafosse A, Maller J, Meslin C, Bonafé A et al. Suicidal 18 behavior is associated with reduced corpus callosum area. Biol Psychiatry 2011; 70: 320–326. 226 Kim B, Oh J, Kim MK, Lee S, Tae WS, Kim CM et al. White matter alterations are associated with suicide attempt in patients with panic disorder. J Affect Disord 2015; 175: 139–146. 227 Thomas LA, Bellis MD De. Pituitary Volumes in Pediatric Maltreatment-Related Posttraumatic Stress Disorder. Biol Psychiatry 2004; 55: 725–758. 228 Caplan R, Siddarth P, Levitt J, Gurbani S, Shields WD, Sankar R. Suicidality and brain volumes in pediatric epilepsy. Epilepsy Behav 2010; 18: 286–290. 229 Ehrlich S, Noam GG, Lyoo IK, Kwon BJ, Clark MA, Renshaw PF. Subanalysis of the location of white matter hyperintensities and their association with suicidality in children and youth. Ann N Y Acad Sci 2003; 1008: 265–268. 230 Ehrlich S, Breeze JL, Hesdorffer DC, Noam GG, Hong X, Alban RL et al. White matter hyperintensities and their association with suicidality in depressed young adults. J Affect Disord 2005; 86: 281–287. 231 Pompili M, Ehrlich S, De Pisa E, Mann JJ, Innamorati M, Cittadini A et al. White matter hyperintensities and their associations with suicidality in patients with major affective disorders. Eur Arch Psychiatry Clin Neurosci 2007; 257: 494–499. 232 Pompili M, Innamorati M, Mann JJ, Oquendo MA, Lester D, Del Casale A et al. Periventricular white matter hyperintensities as predictors of suicide attempts in bipolar disorders and unipolar depression. Prog Neuro-Psychopharmacology Biol Psychiatry 2008; 32: 1501–1507. 233 Ahearn EP, Jamison KR, Steffens DC, Cassidy F, Provenzale JM, Lehman A et al. MRI correlates of suicide attempt history in unipolar depression. Biol Psychiatry 2001; 50: 266–270. 234 Sachs-Ericsson N, Hames JL, Joiner TE, Corsentino E, Rushing NC, Palmer E et al. Differences between suicide attempters and nonattempters in depressed older patients: Depression severity, white-matter lesions, and cognitive functioning. Am J Geriatr Psychiatry 2014; 22: 75–85. 235 Lee SJ, Kim B, Oh D, Kim MK, Kim KH, Bang SY et al. White matter alterations associated with suicide in patients with schizophrenia or schizophreniform disorder. Psychiatry Res - Neuroimaging 2016; 248: 23–29. 236 Audenaert K, Van Laere K, Dumont F, Slegers G, Mertens J, Van Heeringen C et al. Decreased frontal serotonin 5-HT2areceptor binding index in deliberate self-harm patients. Eur J Nucl Med 2001; 28: 175–182. 237 Van Heeringen C, Audenaert K, Van Laere K, Dumont F, Slegers G, Mertens J et al. Prefrontal 5-HT2areceptor binding index, hopelessness and personality characteristics in attempted suicide. J Affect Disord 2003; 74: 149–158. 238 Fountoulakis K, Lacovides A, Fotiou F, Nimatoudis J, Bascialla F, Ioannidou C et al. Neurobiological and psychological correlates of suicidal attempts and thoughts of 19 death in patients with major depression. Neuropsychobiology 2004; 49: 42–52. 239 Henningsson S, Borg J, Lundberg J, Bah J, Lindström M, Ryding E et al. Genetic Variation in Brain-Derived Neurotrophic Factor Is Associated with Serotonin Transporter but Not Serotonin-1A Receptor Availability in Men. Biol Psychiatry 2009; 66: 477–485. 240 Bah J, Lindström M, Westberg L, Mannerås L, Ryding E, Henningsson S et al. Serotonin transporter gene polymorphisms: Effect on serotonin transporter availability in the brain of suicide attempters. Psychiatry Res - Neuroimaging 2008; 162: 221–229. 241 Lindström MB, Ryding E, Bosson P, Ahnlide JA, Rosén I, Träskman-Bendz L. Impulsivity related to brain serotonin transporter binding capacity in suicide attempters. Eur Neuropsychopharmacol 2004; 14: 295–300. 242 Yeh YW, Ho PS, Chen CY, Kuo SC, Liang CS, Ma KH et al. Incongruent reduction of serotonin transporter associated with suicide attempts in patients with major depressive disorder: A positron emission tomography study with 4-[18F]-ADAM. Int J Neuropsychopharmacol 2015; 18: 1–9. 243 Miller JM, Hesselgrave N, Ogden RT, Sullivan GM, Oquendo MA, Mann JJ et al. Positron Emission Tomography Quantification of Serotonin Transporter in Suicide Attempters with Major Depressive Disorder. . Biol Psychiatry 2013; 74: 287–295. 244 Jollant F, Near J, Turecki G, Richard-Devantoy S. Spectroscopy markers of suicidal risk and mental pain in depressed patients. Prog Neuro-Psychopharmacology Biol Psychiatry 2017; 73: 64–71. 245 Rocha MV, Nery-Fernandes F, Guimarães JL, De Castro Quarantini L, De Oliveira IR, Ladeia-Rocha GG et al. Normal metabolic levels in prefrontal cortex in euthymic bipolar I patients with and without suicide attempts. Neural Plast 2015; : 1–9. 246 Cao J, mei Chen J, Kuang L, Ai M, dong Fang W, Gan Y et al. Abnormal regional homogeneity in young adult suicide attempters with no diagnosable psychiatric disorder: A resting state functional magnetic imaging study. Psychiatry Res - Neuroimaging 2015; 231: 95–102. 247 Pan LA, Batezati-Alves SC, Almeida JRC, Segreti A, Akkal D, Hassel S et al. Dissociable patterns of neural activity during response inhibition in depressed adolescents with and without suicidal behavior. J Am Acad Child Adolesc Psychiatry 2011; 50: 602–611. 248 Pan L, Segreti A, Almeida J, Jollant F, Lawrence N, Brenta D et al. Preserved hippocampal function during learning in the context of risk in adolescent suicide attempt. Psychiatry Res Neuroimaging 2013; 211: 112–118. 249 Jollant F, Lawrence NS, Olie E, O’Daly O, Malafosse A, Courtet P et al. Decreased activation of lateral orbitofrontal cortex during risky choices under uncertainty is associated with disadvantageous decision-making and suicidal behavior. Neuroimage 2010; 51: 1275–1281. 250 Cullen KR, Westlund MK, Klimes-Dougan B, Mueller BA, Houri A, Eberly LE et al. 20 Abnormal amygdala resting-state functional connectivity in adolescent depression. JAMA Psychiatry 2014; 71: 1138–1147. 251 Zhang H, Wei X, Tao H, Mwansisya TE, Pu W, He Z et al. Opposite Effective Connectivity in the Posterior Cingulate and Medial Prefrontal Cortex between First- Episode Schizophrenic Patients with Suicide Risk and Healthy Controls. PLoS One 2013; 8: 1–8. 252 Marchand WR, Lee JN, Johnson S, Gale P, Thatcher J. Differences in functional connectivity in major depression versus bipolar II depression. J Affect Disord 2013; 150: 527–532. 253 Tsujii N, Mikawa W, Tsujimoto E, Adachi T, Niwa A, Ono H et al. Reduced left precentral regional responses in patients with major depressive disorder and history of suicide attempts. PLoS One 2017; 12: e0175249. 254 Pu S, Nakagome K, Yamada T, Yokoyama K, Matsumura H, Yamada S et al. Suicidal ideation is associated with reduced prefrontal activation during a verbal fluency task in patients with major depressive disorder. J Affect Disord 2015; 181: 9–17. 21 FIGURE LEGENDS Figure 1: Number of neuroimaging studies on suicidal thoughts and behaviors published in the last two decades. The figure was based on the studies included in this review, calculated separately for studies including adolescents and studies only including adults and divided into separate 4-year time bins for publication date. Figure 2: Overview of brain regions and structural connections included in this review (A) Brain regions that have been most reported in neuroimaging studies investigating structural, functional and molecular brain alterations associated with suicidal thoughts and behaviors, with a subset of regions grouped more broadly into ventral prefrontal cortex, dorsal prefrontal cortex, insula, mesial temporal, subcortical, and posterior regions. (B) White matter tracts implicated in suicidal thoughts and behaviors reported in diffusion tensor imaging studies. DMPFC, dorsomedial prefrontal cortex; dACC, dorsal anterior cingulate cortex; RMPFC, rostromedial prefrontal cortex; mOFC, medial orbitofrontal cortex; vACC, ventral anterior cingulate cortex; PCC, posterior cingulate cortex; Thal, thalamus; VS, ventral striatum; Hippo, hippocampus; Amyg, amygdala; DLPFC, dorsolateral prefrontal cortex; RLPFC, rostrolateral prefrontal cortex; IFG, inferior frontal gyrus; lOFC, lateral orbitofrontal cortex; Put, putamen; Caud, caudate. Figure 3: A tentative brain circuitry model of suicidal thoughts and behaviors. Medial VPFC (ventral ACC, OFC, RPFC), insula, amygdala, hippocampus, lateral temporal regions, posterior midline structures (posterior cingulate cortex and precuneus), dACC, ventral striatum, thalamus and cerebellum contribute to the generation of suicidal ideation through their roles in excessive negative and blunted positive internal states, negative self- referencing, impairments in future thinking and rumination. DPFC (DLPFC and DMPFC), IFG and dACC alterations further exacerbate suicidal thoughts and facilitate suicide behaviors due to their involvement in diminished cognitive control of thought, emotion and behavior and impairments in cognitive flexibility and valuation of different decision options. Alterations in bottom-up and top-down connections between these extended medial VPFC and DPFC/IFG systems may contribute to the transition from suicidal thoughts to behaviors. The dACC and insula may mediate this transition. Dashed lines indicate speculative associations that need further confirmation by future structural and functional connectivity studies. DMPFC, dorsomedial prefrontal cortex; dACC, dorsal anterior cingulate cortex; RPFC, rostral prefrontal cortex; OFC, orbitofrontal cortex; vACC, ventral anterior cingulate cortex; PCC, posterior cingulate cortex; Thal, thalamus; VS, ventral striatum; Hippo, hippocampus; Amyg, amygdala; DLPFC, dorsolateral prefrontal cortex; IFG, inferior frontal gyrus; Put, putamen; Caud, caudate. Table 1. Findings from Structural Imaging Studies of Suicidal Thoughts and Behaviors Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** STRUCTURAL MAGNETIC RESONANCE IMAGING STUDIES OF GRAY AND WHITE MATTER Suicide attempt studies Goodman et al 201187 BPD + MDD 13 SA 13 HC 20 (77) SA: 15.8 (1.1), HC: 16.2 (0.8) ROI volume: OFC, ACC, DMPFC, DLPFC Number of SA associated with  ACC (BA24) volume (combined WM and GM), and WM (not GM) in posterior cingulate (BA 23), however, similar association with BPD symptom severity for both findings. Fradkin et al 201755 MDD 29 SA 29 HC 46 (79) SA: 17.6 (1.82), HC:16.9 (1.72) WB cortical thickness and surface area In HC: motor impulsivity associated with  RMPFC/RLPFC thickness. In SA: motor impulsivity associated with  RMPFC/RLPFC thickness. In SA: non-planning impulsivity associated with  paracentral lobule thickness. No findings medicated vs unmedicated. Cao et al 2016168 DD 35 SA 18 DC, 47 HC 66 (66) SA: 20.63 (3.65), DC: 21.39 (3.05), HC: 20.53 (1.84) WB VBM GMV and WMV (SPM) SA + SI vs DC + SI: No differences between groups Johnston et al 201736 BD 26 SA 42 DC, 45 HC 43 (63) SA: 20.5(3.0), DC:20.6 (3.2), HC:20.8(3.3) WB, VBM GMV (SPM) SA +SI vs DC:  GMV in right medial/lateral OFC (BA11/47), hippocampus, bilateral cerebellum Pan et al 2015115 MDD 28 SA 31 DC, 41 HC ND SA: 16.0 (1.27), DC: 16.06 (1.47), HC: 14.48 (1.84) WB GMV, WMV and cortical thickness SA + SI vs DC:  caudal middle frontal gyrus (BA8) volume, temporal pole (BA38) thickness, parahippocampal gyrus (BA34) volume Peng et al 2014146 MDD 20 SA 18 DC, 28 HC 38 (58) SA: 27.75 (7.21), DC: 31.06 (7.39), HC: 28.61 (5.45) WB VBM GMV (SPM) SA vs DC (SI not reported):  left PCC, which correlated negatively with dysfunctional attitudes. Gosnell et al 201633 MDD, BD, AA, ANX 20 SA 20 DC, 20 HC 28 (47) SA: 28.9 (9.98), DC: 29.25 (11.1), HC: 28.9 (10.0) ROI volume: thalamus, insula, basal ganglia, hippocampus, amygdala, corpus callosum, cortical lobes. SA vs DC:  right precentral gyrus, right IFG, right caudal middle frontal gyrus (DPFC), left precentral lobule and total temporal cortex. No association between SI (regardless of history of SA) and any of the ROIs Soloff et al 2012100 BPD 44 SA 24 DC, 52 HC 76 (63) SA: 29.6 (8.0), DC: 25.9 (5.7), HC: 25.9 (7.2) ROI VBM GMV (SPM), ROIs: IFG, OFC, ACC, middle and superior temporal cortex, insula, hippocampus, parahippocampus, fusiform gyrus, lingual gyrus and amygdala SA vs DC:  insula and larger lingual and middle and superior temporal gyri. In SA: HL associated with  lateral OFC, middle and superior temporal gyri, insula, fusiform gyrus, lingual gyrus and parahippocampal gyrus Monkul et al 2007109 MDD 7 SA 10 DC, 17 HC 34 (100) SA: 31.4 (13.9), DC: 36.5 (7.5), HC: 31.3 (8.3) ROI manual shape tracing, ROIs: OFC, ACC, PCC, amygdala and hippocampus. DLPFC, subgenual PFC, thalamus, temporal lobe, caudate and lateral ventricles in exploratory analyses SA vs DC:  right amygdala Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Besteher et al 201671 SCZ 14 SA 23 DC, 50 HC 46 (53) SA: 34.4 (12.1), DC: 28.8 (9.7), HC: 29.5 (7.9) WB cortical thickness and mean curvature SA vs DC: SA  in right DLPFC and superior and middle temporal gyri, temporopolar cortex, and insula Giakoumatos et al 201437 SCZ, SZA or BD-P 148 SA (97 HL, 51 LL) 341 DC, 262 HC 387 (52) SA-HL: 35.6 (11.7), SA-LL: 36.9 (12.2), DC: 35.9 (13.3), HC: 38.1 (12.5) WB GMV, cortical surface area and thickness SA vs DC:  GMV in bilateral superior and middle frontal gyri (DLPFC), and inferior and superior temporal gyri, left superior parietal and supramarginal cortex, and right insula and thalamus. High (vs. low) lethality:  GMV in left lingual area and right cuneus across all attempters; left dorsal ACC (BA32), left inferior parietal cortex, left inferior temporal gyrus, and right middle temporal gyrus in BD-P; left lingual gyrus, bilateral pericalcarine, right cuneus and right lateral occipital cortex in SZ; left middle frontal gyrus (DLPFC) in SZA. Rüsch et al 2008219 SCZ 10 SA 45 DC, 55 HC 42 (38) SA: 30.3 (6.5): DC 37.3 (11.6) WB VBM GMV and WMV (SPM) SA vs DC:  WMV in bilateral posterior lateral OFC and IFG. No GMV differences Matsuo et al 2010220 BD 10 SA 10 DC, 27 HC 47 (100) SA: 36.2 (10.1), DC:44.2 (12.5), HC: 36.9 (13.8) ROI volume, manual shape tracing of CC genu, anterior body, posterior body, isthmus and splenium SA vs DC: no significant differences. In SA: impulsivity associated with  anterior corpus callosum genu Lijffijt et al 2014113 BD 51 SA 42 DC, 45 HC 138 (100) SA: 36.6 (10.7), DC: 41.1 (11.3) ROI volume, ROIs: superior frontal gyrus, rostral and caudal middle frontal gyrus, frontal pole, IFG, medial and lateral OFC, rostral and caudal ACC SA vs DC: no differences. Lower PFC GMV only in SA with previous hospitalization Soloff et al 201488 BPD 51 SA (16 HL, 35 LL) 41 (80) SA-HL: 36.1 (9.2): SA-LL: 27.4 (5.9). ROI volume VBM of GMV (SPM), ROIs: OFC, ACC, middle and superior temp gyrus, insula, (para-)hippocampus, lingual gyrus, amygdala HL vs LL SA:  GMV in bilateral middle and superior temporal gyri, left lingual gyrus, bilateral lateral OFC, right insula, bilateral fusiform gyrus, right parahippocampus, left ventral and dorsal ACC and left hippocampus. In HL SA: aggression associated with  ventral and dorsal ACC, lateral OFC, middle and superior temporal gyri and right insula, and impulsivity associated with  right middle and superior temporal gyri and  insula. In LL SA: impulsivity associated with  right middle and superior temporal gyri, bilateral insula, bilateral lingual gyrus, ventral ACC, fusiform gyrus, lateral OFC, hippocampus and amygdala Aguilar et al 2008142 SCZ 13 SA 24 DC 0 (0) SA: 37.1 (11.0), DC: 42.7 (10.2) WB VBM GMV (SPM) SA vs DC:  medial OFC and superior temporal gyrus Ding et al 201532 Past MDD/BD 67 SA 82 DC, 82 HC 98 (42) SA: 39.2 (10.6), DC: 39.4 (9.7),HC: 37.8 (8.1) WB and ROI VBM GMV (SPM) and cortical thickness and surface area, ROIs: OFC, VLPFC, VMPFC (including ACC), and DLPFC SA vs DC:  OFC (BA47, lateral part of BA11) in exploratory WB analysis. In SA: lethality last SA associated with  right DPFC BA8/9/46), OFC, left VLPFC (BA44/45). Number SA associated with  right DPFC and left OFC. No associations with SI. Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Gifuni et al 2017221 Past MDD/BD 61 SA 75 DC, 73 HC 120 (47) SA: 38.3 (10.7), DC: 38.4 (9.1), HC: 39.2 (7.0) ROI volume SBM, ROI: corpus callosum SA vs DC: no differences. No correlation corpus callosum volume and SI, age at first SA and number of SA Gifuni et al 2016112 Past DD/BD 73 SA 89 DC, 91 HC 120 (47) SA: 39.2 (10.6), DC: 39.4 (9.5), HC: 38.3 (8.2) ROI volume, ROIs: amygdala, hippocampus, caudate, globus pallidus, putamen, nucleus accumbens, ventral diencephalon and thalamus SA vs DC: no differences. In SA: lethality SA associated with  left and right nucleus accumbens Harenski et al 2017222 SCZ, SZA, BD- P, or MDD-P 18 SA 18 DC, 59 HC, 26 HC 0 (0) SA: 38.9 (11.73), DC: 40.2 (10.23), HC: 32.5 (11.16), CHC: 33.0 (9.49) WB and ROI VBM GMV (SPM), ROIs: posterior superior temporal cortex, temporal poles and medial PFC SA vs DC:  left and right temporal pole Nery- Fernandes et al 2012194 BD 19 SA 21 DC, 22 HC 41 (66) SA: 39.8 (11.4), DC: 42.0 (8.6), HC: 37.7 (13.5) ROI volume VBM GMV (SPM), ROI: corpus callosum SA vs DC: no differences Vang et al 2010130 MDD, AD 7 SA 6 HC ND SA: 40 (11.83), matched HC, no details ROI volume, ROIs: subcortical structures SA vs HC:  globus pallidus and caudate, and correlated with 5-HTT binding. In SA: non-impulsive temperament associated with  globus pallidus GMV Baldaçara et al 2011223 BD 20 SA 20 DC, 22 HC 41 (66) SA: 39.94 (11.2), DC: 41.9 (8.9), HC: 37.7 (13.6) ROI volume VBM GMV and WMV (SPM), ROIs: cerebellum SA vs DC: no differences total brain volume or cerebellar volume Wagner et al 201134 MDD 10 with SA and/or first- degree relative with SA 15 DC, 30 HC 50 (83) SA: 41.0 (12.5): DC: 34.1 (10.5), HC: 35.1 (10.4). WB VBM GMV(SPM) SA vs DC:  rostral ACC (BA24) and right caudate Wagner et al 201258 MDD 10 with SA and/or first- degree relative with SA 15 DC, 30 HC 50 (83) SA: 41.0 (12.5), DC: 34.1 (10.5), HC: 35.1 (10.4) WB cortical thickness SA vs DC:  VLPFC (BA47), DLPFC (BA46) and dorsal ACC (BA32). Patients with own versus relative with SA: no differences in dorsal ACC, VLPFC, DLPFC. Duarte et al 2017101 BD 20 SA 19 DC, 20 HC 34 (57) SA: 41.10 (12.64): DC: 42.26 (11.70), HC: 37.40 (10.20) WB and ROI volume VBM GMV (SPM), ROIs: OFC, DLPFC (including IFG), ACC, amygdala, hippocampus, thalamus and insula SA vs DC:  right rostral ACC (BA24). In SA: HL SA was associated with  insula, LL SA was associated with  OFC (BA47) Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Benedetti et al 2014224 BD 32 SA 104 DC 93 (68) SA (5-HTTLPR l/l): 41.4 (10.7), SA (5-HTTLPR *s): 46.4 (12.8), DC (5-HTTLPR l/l): 48.5 (10.4), DC (5-HTTLPR *s): 46.8 (12.6) WB GMV (SPM) SA vs DC: no differences. Lee et al 2016150 MDD 19 SA 19 DC, 20 HC 41 (73) SA: 42.0 (10.8), DC: 41.1 (15.2) ROI volume GMV (SPM), ROIs not specified SA vs DC:  right cerebellum and left angular gyrus Spoletini et al 2011110 SCZ 14 SA 36 DC, 50 HC 35 (39) SA: 42.9 (11.3), DC: 39.8 (11.4): HC: 40.0 (16.6) ROI volume GMV (FSL), ROIs: lateral ventricles, thalamus, caudate, putamen, pallidum, hippocampus, amygdala, nucleus accumbens SA vs DC:  right amygdala Benedetti et al 201135 BD 19 SA (with/without Lithium) 38 DC (with/without Lithium) 38 (67) SA-L-: 43.6 (10.4), SA-L+: 45.6 (11.3), DC- L-: 45.9 (10.5), DC-L+: 46.2 (13.3) WB VBM GMV (SPM) SA vs DC:  DPFC (BA6/8/9), RLPFC (BA10), OFC (BA11/47), dorsal ACC (BA32), parietal and occipital cortex and  in bilateral superior temporal gyrus. SA with lithium vs without:  DPFC (BA6/8), OFC (BA11/47), ACC (BA24/32), parietal and occipital cortex and  in bilateral superior temporal gyrus Colle et al 2015114 MDD 24 SA 39 DC 39 (62) SA: 44.2 (11.9), DC: 47.7 (12.6) ROI volume GMV (SACHA, automatic segmentation), ROI: hippocampus. SA vs DC:  hippocampus. No difference between SA in last month versus SA >1 month ago Dombrovski et al 2012129 MDD 13 SA 20 DC, 19 HC 30 (58) SA: 66.0 (6.4), DC: 67.7 (7.0), HC: 70.5 (7.5) ROI voxel count basal ganglia (caudate, putamen, pallidum) SA vs DC:  putamen, associative and ventral striatum voxel count. In SA: delay discounting associated with  putamen voxel count Cyprien et al 2011225 MDD, ANX, BD 21 SA or SI 234 DC, 180 HC 222 (51) SA: 72.2 (4.3), DC: 71.0 (3.8), HC: 71.0 (3.8) ROI volume manual shape tracing, ROI: corpus callosum SA vs DC:  posterior third of corpus callosum Hwang et al 201070 MDD 27 SA 43 DC, 26 HC 0 (0) SA:79.1 (5.6), DC:79.6 (5.1), HC: 79.5 (4.3) WB VBM GMV and WMV (SPM) SA vs DC:  GMV in DPFC (BA6/8/9/46), precentral gyrus, postcentral gyrus, superior parietal lobe, inferior parietal lobe, cuneus, superior temporal gyrus, insula, cerebellum, midbrain,  WMV in DPFC (BA6/8/9/46), precentral and postcentral gyrus, inferior parietal lobe, precuneus, occipital lobe, external capsule, cerebellum Lopez-Larson et al 2013134 TBI 19 SA 40 DC, 15 HC 0 (0) 18-55 ROI volume, ROI: thalamus SA vs DC:  right thalamus Jia et al 2010165 MDD 16 SA 36 DC, 52 HC 55 (53) SA: 34.2 (13.7), DC: 34.7 (12.5), HC: 37.1 (16.0) WB VBM GMV and WMV (SPM) SA vs DC: no differences Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Kim et al 2015226 PD 12 SA 25 DC 23 (64) 16-60 WB VBM GMV and WMV (SPM) SA vs DC: no differences Rentería et al 2017111 MDD 153 SA or SI+plan, 298 SI-plan 650 DC, 1996 HC ND SA+SI: 21.10- 53.8 (across 7 samples), DC: 22.9-54.8, HC: 22.9-55.4. ROI volume, ROIs: nucleus accumbens, amygdala, caudate, hippocampus, pallidum, putamen, thalamus, ICV SA+SI with plan vs DC: no differences. SI vs DC: no differences Suicidal ideation studies Thomas et al 2004227 PTSD + Childhood maltreatment 47 SI, 17 SA 14 DC, 121 HC 89 (49) DC: 11.71 (2.6): HC: 11.74 (2.5) ROI volume manual shape tracing (IMAGE), ROI: pituitary SI (with SA) vs DC:  pituitary Taylor et al 201538 MDD 21 SI, including 10 with past SA 53 DC, 91 HC 108 (65) SI: 33.5 (9.1), DC:37.5 (8.9), HC: 29.9 (9.1) WB and ROI volume GMV and cortical thickness, ROIs: OFC, cingulate cortex, insula, amygdala, parahippocampus, thalamus, basal ganglia SI+SA vs DC:  cortical thickness of left insula, left caudal middle frontal gyrus (DLPFC), left superior parietal cortex, left superior temporal gyrus. No GMV differences Caplan et al 2010228 Epilepsy & MDD, ANX, ADHD 11 SI (No past SA) 40 DC 28 (55) SI: 11.04 (2.06), DC:9.43 (2.07) ROI volume manual shape tracing, ROIs: middle frontal gyrus, superior frontal gyrus, OFC, temporal lobe SI vs DC:  right orbital frontal gyrus WMV and  left temporal lobe GMV MAGNETIC RESONANCE IMAGING STUDIES OF WHITE MATTER HYPERINTENSITIES Suicide attempt studies Ehrlich et al 2003229 MDD, BD, PsD, conduct/ADHD 43 SA 110 DC 39 (26) Entire sample: 14.6 (3.4). No further details provided. WMH (Modified version of Coffey scale) SA vs DC:  DWMHs in parietal lobes. All SA subjects had lesions in right posterior parietal lobe Ehrlich et al 2004198 MDD, BD, PsD, conduct/ADHD 43 SA 110 DC 41 (27) Entire sample: 14.6 (3.4). No further details provided. WMH (Modified version of Coffey scale)  WMH associated with past SA, driven by PVH. SI not associated with WMH Ehrlich et al 2005230 MDD 62 SA 40 DC 68 (67) Entire sample: 26.7 (5.5). No further details provided. WMH (Modified version of Fazekas scale) SA vs DC:  PVH, not DWMH. No association with SI Pompili et al 2007231 MDD, BD 29 SA 26 DC 41 (63) SA: 42.2 (13.5), DC: 44.6 (14.0) WMH (Modified version of Fazekas scale) SA vs DC:  WMH. SI was not associated with WMH Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Pompili et al 2008232 MDD, BD 44 SA 55 DC 57 (58) SA: 45.57 (16.10), DC: 47327(14.54) WMH (Modified version of Fazekas scale) SA vs DC:  PVH, no difference DWMH Ahearn et al 2001233 MDD 20 SA 20 DC 17 (85) SA: 66.0 (5.8), DC: 66.4 (5.7) WMH (Coffey and Boyko scales) SA vs DC:  subcortical GM hyperintensities, and trend towards more PVH Sachs- Ericsson et al 2014234 MDD 23 SA 223 DC 149 (67) SA: 66.74 (6.6), DC: 69.8 (7.5) WMH (Duke Neuropsychiatric Imaging Research Laboratory modified version of MrX software) SA vs DC:  WM lesions in the left hemisphere.  increase over time in bilateral WMH in SA, which was predicted by the number of depressive episodes. DIFFUSION TENSOR IMAGING STUDIES Suicide attempt studies Johnston et al 201736 BD 26 SA 42 DC, 45 HC 43 (63) 14-25 WB FA maps (SPM) SA vs DC:  FA in left uncinate fasciculus, right uncinate fasciculus and right cerebellum Lischke et al 2017158 BPD 13 SA 8 DC, 20 HC 41 (100) 18-45 Tractography based FA and MD with seeds in genu, splenium and body of corpus callosum SA vs DC: no differences. Number of attempts associated with  FA and MD in splenium and FA in the genu Lee et al 2016235 SCZ 15 SA 41 DC 41 (73) 18-60 WB FA, MD, AD and RD maps (FSL) SA vs DC:  FA in left corona radiata (anterior, superior, posterior), superior longitudinal fasciculus, posterior limb and retrolenticular part of internal capsule, external capsule, posterior thalamic radiation, sagittal stratum (including inferior longitudinal fasciculus and inferior fronto-occipital fasciculus),  AD in retrolenticular part of internal capsule, posterior thalamic radiation and sagittal stratum. No differences in MD and RD. Kim et al 2015226 PD 12 SA 25 DC 23 (64) 16-60 FA, MD, AD and RD maps (FSL), Tracts: corona radiata, inferior longitudinal fasciculus, inferiorfronto-occipital fasciculus, superior longitudinal fasciculus, posterior thalamic radiation, internal capsule, splenium of corpus callosum SA vs DC:  FA in posterior and superior corona radiata, sagittal stratum (including inferior longitudinal fasciculus and inferior fronto-occipital fasciculus), superior longitudinal fasciculus, posterior thalamic radiation, retrolenticular part of internal capsule and splenium of the corpus callosum. No differences in MD, RD and AD. In SA: positive correlation between suicidal ideation and FA of the right retrolenticular part of internal capsule and right and left posterior thalamic radiation. In DC: positive correlation between suicidal ideation and FA of splenium, right retrolenticular part of internal capsule and left posterior thalamic radiation Mahon et al 2012156 BD 14 SA 15 DC, 15 HC 18 (41) SA: 33.3 (14.1), DC: 36.5 (12.3), HC: 33.7 (12.6) WB FA maps SA vs DC:  FA in white matter tract in medial VPFC. In SA: medial VPFC FA negatively correlated with motor impulsivity Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Olvet et al 2014169 MDD 13 SA 39 DC, 46 HC 52 (53) 18-65 ROI FA and ADC maps (FSL), ROIs: mOFC, DMPFC, rACC, dACC SA vs DC:  FA in DMPFC. No difference in AD Jia et al 2010165 MDD 16 SA 36 DC, 52 HC 55 (53) SA: 34.2 (13.7), DC: 34.7 (12.5), HC: 37.1 (16.0) WB and ROI analysis of FA, MD and RD maps (DTIstudio), ROIs: bilateral lentiform nucleus, bilateral hippocampus, and bilateral thalamus SA vs DC:  FA and AD in the left anterior limb of internal capsule and  FA and  RD in right lentiform nucleus Jia et al 2014166 MDD 23 SA 40 DC, 46 HC 59 (54) SA: 36.3 (14.5, DC: 34.0 (14.5), HC: 33.3 (11.4) Tractography based FA with seed in left anterior limb of the internal capsule SA vs DC:  percentage of projecting fibers connecting the anterior limb of the internal capsule to the left OFC and left thalamus Lopez-Larson et al 2013134 TBI 19 SA 40 DC, 15 HC 0 (0) 18-55 ROI FA maps (FSL), ROI: anterior thalamic radiation SA vs DC:  FA in bilateral anterior thalamic radiation. Positive correlation impulsivity and FA in right anterior thalamic radiation Cyprien et al 2016157 BD, MDD 45 SA 46 DC, 30 HC 121 (100) 18-50 ROI FA, MD, RD, AD maps (FSL), ROIs: genu, body and splenium of corpus callosum SA vs HC: no differences that survived multiple comparison correction. Number of attempts associated with  FA in genu, body and splenium of corpus callosum. FA in splenium negatively correlated with suicidal intent. Bijttebier et al 2015164 past MDD 13 SA 15 DC, 17 HC 32 (72) 18-65 Tractography combined with network-based statistics SA vs DC:  structural connectivity in network including medial VPFC, temporal gyrus, precuneus, cuneus, parietal cortex, amygdala, hippocampus, occipital regions. Decreased connectivity between left olfactory cortex and left ACC Suicidal ideation studies Myung et al 2016163 MDD 24 SI 25 DC, 31 HC 52 (65) 18-62 Tractography combined with network-based statistics and graph analysis SA vs DC:  structural connectivity in left hemisphere network including striatal regions, frontal regions (DLPFC, IFG, lateral OFC and RMPFC), lateral occipital and superior parietal regions. Betweenness centrality of left rostral middle frontal gyrus (DLPFC) positively correlated with suicidal ideation. Participation coefficient of left rostral middle frontal gyrus (DLPFC) positively correlated with impulsivity Taylor et al 201538 MDD 21 SI, including 10 with past SA 53 DC, 91 HC 108 (65) 20-50 ROI FA and MD maps (FSL), ROIs: internal capsule, thalamic radiation, cingulum bundle, corpus callosum, uncinate fasciculus SA vs DC:  RD and  FA in the corona radiata, the hippocampal region of the cingulum and the anterior thalamic radiation Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Symbols & Abbreviations: *Percentages are rounded to the nearest whole number; **Results are reported for SA or SI in comparison with diagnostic controls. If no diagnostic controls were included in the study, results based on SA or SI compared to healthy controls are reported; AA: alcohol abuse; ACC: anterior cingulate cortex; AD: adjustment disorder; ADHD: attention deficit hyperactivity disorder; ANX: anxiety disorder; BA: Broadman’s Area; BD: bipolar disorder; BD-P: bipolar disorder with psychotic symptoms; BPD: borderline personality disorder; DC: diagnostic controls; DD: depressive disorder; DLPFC: dorsolateral prefrontal cortex; DMPFC: dorsomedial prefrontal cortex; DWMH: deep white matter hyperintensities; FA: fractional anisotropy; GM: grey matter; GMV: grey matter volume; HC: healthy controls; HL: high lethality; IFG: inferior frontal gyrus; LL: low lethality; MD: mood disorder; MDD: major depressive disorder; MDD-P: major depressive disorder with psychotic features; ND: not detailed; OFC: orbitofrontal cortex; PCC: posterior cingulate cortex; PD: personality disorder; PFC: prefrontal cortex; PsD: psychosis; PVH: periventricular hyperintensities; RLPFC: rostrolateral prefrontal cortex; RMPFC: rostromedial prefrontal cortex; ROI: region of interest; SA: suicide attempt; SBM: surface based morphometry; SCZ: schizophrenia; SI: suicidal ideation; SUD: substance use disorder; SZA: schizoaffective disorder; SPM: Statistical Parametric Mapping toolbox; TBI: traumatic brain injury; VBM: voxel based morphometry; VLPFC: ventrolateral prefrontal cortex; VMPFC: ventromedial prefrontal cortex; WB: whole brain; WM: white matter; WMH: white matter hyperintensities; WMV: white matter volume; 5-HTTLPR l/l: serotonin transporter long/long genotype; 5-HTTLPR*s: serotonin transporter s allele carriers genotype. Table 2. Findings from Molecular Imaging Studies of Suicidal Thoughts and Behaviors Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** SINGLE PHOTON EMISSION TOMOGRAPHY STUDIES Suicide attempt studies Audenaert et al 2001236 MDD, AD, PsD 9 SA 12 HC 8 (38) 19-48 123I-5-I-R91150 for 5-HT2a receptors in PFC SA vs HC:  binding potential of 5-HT2a receptors in PFC Audenaert et al 2002143 MDD 20 SA 20 HC 24 (60) 19-50 99mTc-Ethyl Cystine Dimer rCBF SPECT during letter and category fluency SA vs HC:  perfusion in IFG, ACC, temporal gyrus, hypothalamus during verbal fluency task van Heeringen et al 2003237 MDD, AD, PsD 9 SA 13 HC ND 19-47 123I-5-I-R91150 for 5-HT2a receptors in PFC SA vs HC:  binding potential of 5-HT2a receptors in PFC Amen et al 2009102 MDD 12 SA 12 DC, 12 HC 3 (8) 19-64 99mTc HMPAO SPECT to assess rCBF SA vs DC:  rCBF in subgenual ACC,  rCBF in right insula, dorsal ACC Willeumier et al 201182 MD 21 SA 36 DC, 27 HC 5 (24) 15-66 99mTc HMPAO SPECT to assess rCBF SA vs DC:  rCBF in frontal, temporal and parietal regions Fountoulakis et al 2004238 MDD 13 SA, 10 SI 33 DC ND 21-60 99mTc HMPAO SPECT to assess rCBF SA vs DC: no differences. SI vs no-SI: no differences Henningsson et al 2009239 MDD, PD 9 SA 9 HC ND 23-67 123I-β-CIT for 5-HTT binding potential, assessment of Val66Met polymorphisms Within SA: carriers of the Val/Val genotype of Val66Met had  5HTT binding potential in the parietal cortex and in the occipital lobes Bah et al 2008240 MDD, AD, PD 9 SA 9 HC 0 (0) 23-67 123I-β-CIT for 5-HTT binding potential, assessment of SLC6A4 polymorphisms SA vs HC: no differences. In SA: presence of S-allele of 5- HTTLPR genotype associated with lower 5-HTT binding potential in frontal, parietal and occipital cortex Lindström et al 2004241 MDD, AD, PD 12 SA 12 HC 4 (17) 23-67 123I-β-CIT methods to separate 5-HTT and DAT uptake SA vs HC: no differences in 5-HTT or DAT binding. In SA: impulsivity associated with  whole brain 5-HTT binding Ryding et al 2006132 MDD, AD, ANX, PD 12 SA 12 HC 4 (17) 23-67 123I-β-CIT methods to separate 5-HTT and DAT uptake SA vs HC: no differences in 5-HTT or DAT binding. In SA: impulsivity associated with  5-HTT binding potential in inferior/orbital frontal cortex, temporal regions, midbrain, thalamus, basal ganglia Vang et al. 2010130 MDD, AD 7 SA 6 HC 3 (23) SA: 40 (11.83), matched HC, no details 123I-β-CIT methods to separate 5-HTT and DAT uptake SA vs HC: not reported. In SA: significant negative correlation between 5HTT binding and globus pallidus volume Suicidal ideation studies Fountoulakis et al 2004238 MDD 13 SA, 10 SI 33 DC ND 21-60 99mTc HMPAO SPECT to assess rCBF SI vs DC: no differences. SI vs no-SI: no differences Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** POSITRON EMISSION TOMOGRAPHY STUDIES Suicide attempt studies Soloff et al 2003196 BPD 13 SA 9 HC 22 (100) 18-49 [18F]FDG PET during rest SA vs HC:  rCMRglu in bilateral medial OFC Yeh et al 2015133 MDD 5 SA 5 DC, 10 HC 0 (0) 20-25 4-[18F]-ADAM for SERT availability SA vs DC:  SERT binding potential in the midbrain, thalamus, striatum and PFC. Suicidal ideation associated with  SERT binding potential in the same 4 regions Soloff et al 2014197 BPD 21 SA 12 DC, 27 HC 32 (53) BPD: 27.5 (7.2), HC: 28.8 (8.2) [18F]altanserin for 5-HT2a receptor binding potential SA vs DC:  binding potential of 5-HT2a receptors in the occipital cortex in females only Soloff et al 2007116 BPD 12 SA 2 DC, 11 HC 25 (100) 19-46 [18F]altanserin for 5-HT2a receptor binding potential SA vs HC:  binding potential of 5-HT2a receptors in the hippocampus, medial temporal cortex and occipital cortex. No associations with number of attempts Cannon et al 200646 BD 8 SA 10 DC, 37 HC 36 (65) BD: 30 (9), HC: 32 (9) [11C]DASB for 5-HTT binding potential SA vs DC:  5-HTT binding in the midbrain and  in the rostral ACC Oquendo et al 200384 MDD 25 SA (9 LL, 16 HL) 15 (60) LL: 30.4 (8.7), HL: 42.9 (10.4) [18F]FDG PET, fenfluramine vs. placebo challenge HL vs LL SA:  rCMRglu in ACC (BA24/32), IFG (BA44) and DPFC (BA6/8/9), more pronounced after fenfluramine challenge. Lower VMPFC rCMRglu associated with lower impulsivity, higher suicidal intent and higher lethality Sullivan et al 201541 MDD 29 SA 62 DC 59 (65) 18-65 [11C]WAY-100635 for 5-HT1a receptor binding potential SA vs DC: no differences in 5-HT1a receptor binding. High vs Low lethality SA:  5-HT1a receptor binding potential in the raphe nuclei. Positive association 5-HT1a receptor binding potential in the raphe nuclei and suicidal intent. Positive association 5-HT1a receptor binding potential in the raphe nuclei and PFC and suicidal ideation Miller et al 201644 BD, MDD 11 SA 6 DC, 31 HC 29 (60) 21 - 61 [11C]DASB for 5-HTT binding potential SA vs DC: no differences in 5-HTT binding Yeh et al 2015242 MDD 8 SA 9 DC, 17 HC 18 (53) 20-65 4-[18F]-ADAM for SERT availability SA vs DC: no differences in SERT binding in individual regions, but  PFC/midbrain SERT binding ratio. Suicide intent positively associated with PFC/midbrain SERT binding ratio Leyton et al 200640 MD, PD, SUD (all HL) 10 SA 16 HC 7 (28) SA: 37.7 (6.4), HC: 35.5 (12.0) Alpha-11C-methyl-L-tryptophan trapping as index of 5-HT synthesis SA vs HC:  5-HT synthesis in lateral and medial OFC extending into VMPFC,  5-HT synthesis in thalamus, paracentral lobule, occipital cortex and hippocampus. Negative correlation suicide intent and serotonin synthesis in lateral OFC and VMPFC Parsey et al 200643 MDD 9 SA 16 DC, 43 HC 49 (69) MDD: 38.0 (13.4), HC: 38.8 (15.9) [11C]McN 5652 for 5-HTT binding potential SA vs DC: no differences in SERT binding Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Miller et al 2013243 MDD 15 SA 36 DC, 32 HC 41 (49) SA: 38.5 (11.5), DC: 41.0 (10.5), HC: 32.6 (11.3) [11C]DASB for 5-HTT binding potential SA vs DC:  5-HTT binding potential in midbrain Nye et al 2013131 MDD 11 SA 10 HC 8 (38) SA: 38.5 (13.6), HC: 21.3 (2.4) [11C]ZIENT for SERT binding potential SA vs HC:  5-HTT in the midbrain/pons and putamen Mann et al 201845 MDD 8 SA 8 DC, 8 HC ND 21-53 [11C]WAY-100635 for 5-HT1a receptor binding potential, [18F]altanserin for 5-HT2a receptor binding potential SA vs DC: no difference in 5-HT1a and 5-HT2a receptor binding Oquendo et al 201642 MDD 51 past SA, 15 future SA 49 DC 61 (61) 18-65 [11C]WAY-100635 for 5-HT1a receptor binding potential, [11C]DASB for 5-HTT binding potential Future attempt versus no future attempt: no differences in 5-HTT binding. Higher lethality of future attempts associated with  5-HT1a receptor binding potential in insula, DPFC, ACC and raphe nuclei. SI at follow up associated with  5-HT1a receptor binding in raphe nuclei, amygdala, hippocampus, parahippocampul gyrus, temporal lobe, ACC, DPFC, medial PFC, OFC, insula, occipital lobe, parietal lobe Sublette et al 201383 MDD, BD 13 SA 16 DC 19 (66) SA: 36.0 (11.5), DC: 42.2 (13.0) [18F]FDG PET, fenfluramine vs. placebo challenge SA vs DC:  rCMRglu in right DLPFC, more pronounced after fenfluramine challenge, and  rCMRglu in VMPFC, not more pronounced after fenfluramine. SI negatively correlated with rCMRglu in DLPFC Suicidal ideation studies Holmes et al 201890 MDD 9 SI 5 DC, 13 HC 13 (48) MDD: 31 (12), HC: 33 (11) [11C](R)-PK11195 for TSPO availability index of neuroinflammation) SI vs no-SI:  TSPO availability in ACC and insula Kolla et al 201689 BPD 28 DC, 14 HC 56 (100) 18-51 [11C ]Harmine for MAO-A VT (index of MAO-A density) Positive association MAO-A VT in PFC and ACC with SI, but also with depressive symptom scores Van Heeringen et al 201785 MDD 17 SI + plan, 11 SI 12 DC, 20 HC 38 (63) SI+plans: 46.1 (10.9), SI: 42.6 (11.6), DC: 51.2 (6.5), HC: 43.8 (13.1) [18F]FDG PET during rest SI+plans vs SI:  rCMRglu in RLPFC and inferior parietal lobe Ballard et al 2015151 MDD 12 SI 8 DC 6 (30) MDD: 48 (12) [18F]FDG PET during rest, ketamine challenge Baseline SI associated with  rCMRglu in infralimbic cortex. Ketamine induced reductions in SI associated with reductions in rCMRglu in infralimbic cortex and increases in rCMRglu in cluster including lingual gyrus, occipital cortex and cerebellum Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** MAGNETIC RESONANCE SPECTROSCOPY STUDIES Suicide attempt studies Jollant et al 2017244 past MDD 15 SA 10 DC, 33 HC 35 (60) 15 -55 Proton MRS;. Metabolites: glutamate, glutamine, N- acetylaspartate, myo-inositol, aspartate, glutathione, GABA, N- acetylaspartylglutamate, total choline. ROI: right dorsal PFC SA vs DC: no significant differences. Choline levels positively correlated with current suicidal ideation Prescot et al 201891 history or current MDD, PTSD and/or SUD 57 with SA or SI 24 DC 16 (20) SA+SI: 37.2 (9.1), DC: 36.2 (9.7) Proton Magnetic Resonance Spectroscopy. Metabolites: GABA, N- acetylaspartylglutamate, glutamine, glutamate, creatine. ROI: dorsal ACC SA vs DC: no significant differences. In females only,  GABA in SA+SI compared to DC but no longer significant after correcting for age differences Rocha et al 2015245 past BD 19 SA 21 DC, 22 HC 41 (66) SA: 39.8 (11.4), DC: 42.0 (8.6), HCL 37.7 (13.5) Proton MRS; Metabolites: N- acetylaspartate, choline, creatine, myo-inositol. ROI: medial OFC SA vs DC: no significant differences Suicidal ideation studies Gabbay et al 2017195 MDD 44 DC, 36 HC 46 (50) 12-21 Proton MRS; Metabolites: GABA and GLX. ROI: rostral ACC No correlation between GABA and GLX levels in rostral ACC and SI Symbols & Abbreviations: *Percentages are rounded to the nearest whole number; **Results are reported for SA or SI in comparison with diagnostic controls. If no diagnostic controls were included in the study, results based on SA or SI compared to healthy controls are reported; ACC: anterior cingulate cortex; AD: adjustment disorder; ANX: anxiety disorder; BD: bipolar disorder; BPD: borderline personality disorder; DAT: dopamine transporter; DC: diagnostic controls; DLPFC: dorsolateral prefrontal cortex; DPFC: dorsolateral prefrontal cortex; FDG: fludeoxyglucose; GABA: gamma- aminobutyric acid; GLX: glutamate + glutamine; HC: healthy controls; HL: high lethality; IFG: inferior frontal gyrus; LL: low lethality; MAO-A: monoamine oxidase A; MDD: major depressive disorder; MRS: magnetic resonance spectroscopy; ND: not detailed; OFC: orbitofrontal cortex; PET: positron emission tomography; PD: panic disorder; PFC: prefrontal cortex; PsD: psychotic disorder; rCBF: regional cerebral blood flow; rCMRglu: regional cerebral metabolic rate for glucose; RLPFC: rostrolateral prefrontal cortex; SA: suicide attempt; SERT: serotonin transporter; SI: suicidal ideation; SP: social phobia; SPECT: single photon emission computed tomography; SUD: substance use disorder; TSPO: translocator protein, VMPFC: ventromedial prefrontal cortex; 5-HT: serotonin, 5-HTT: serotonin transporter Table 3. Findings from Functional Imaging Studies of Suicidal Thoughts and Behaviors Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** FUNCTIONAL MAGNETIC RESONANCE IMAGING STUDIES Suicide attempt studies Resting State fMRI Cao et al 2016168 DD 35 SA 18 DC, 47 HC 66 (66) SA: 20.63 (3.65), DC: 21.39 (3.05), HC: 20.53 (1.84) WB fractional zALFF SA vs DC:  zALFF in right superior temporal gyrus, left middle temporal gyrus, left middle occipital gyrus, left angular gyrus,  zALFF in left RLPFC. In SA: negative correlation impulsivity and zALFF in left RLPFC Cao et al 2015246 No DX 19 SA 20 HC 22 (56) SA: 19.8 (1.6), HC: 20.3 (1.7) WB ReHo SA vs HC:  ReHo in left fusiform gyrus, lateral OFC (BA47), hippocampus, right angular gyrus, bilateral parahippocampal gyrus, DLPFC (BA46) and cerebellum,  ReHo in right inferior parietal lobe, left precuneus, right medial OFC (BA11) Zhang et al 2016154 DD 35 SA 18 DC, 47 HC 66 (66) SA: 20.63 (3.65), DC: 21.26 (3.02), HC: 20.48 (1.86) ICA of DMN SA vs DC:  FC in the right precuneus,  FC in the left lingual gyrus and left cerebellum Kang et al 2017160 MDD 19 SA 19 DC 20 (53) SA: 42.0 (10.8), DC: 41.1 (15.2) Amygdala seed-based FC SA vs DC:  FC left amygdala with right insula and left OFC (BA11),  FC right amygdala with left middle temporal gyrus. In SA: positive correlation between SI and right amygdala FC with right parahippocampal gyrus Cognitive control Pan et al 2011247 MDD 15 SA 15 DC, 14 HC 25 (57) 12-17 Go-no-go response inhibition, WB SA vs DC:  dorsal ACC and insula activation during response inhibition, driven by greater activity in DC compared to both SA and HC (i.e., no evidence for abnormal response inhibition circuitry in SA) Richard- Devantoy et al 2016137 MDD 25 SA 22 DC, 27 HC 47 (61) 18-55 Go-no-go response inhibition, WB SA vs DC: no significant differences. Suicidal intent was positively associated with thalamus activity during response inhibition Minzenberg et al 201454 SCZ 8 SA+SI, 10 SI only 17 DC 10 (26) 18-50 Continuous performance, ROI frontal cortex SA+SI vs SI:  goal-related left dorsal premotor cortex (BA6) activity. SI vs no-SI:  goal-related ventral ACC, VMPFC, VLPFC, DLPFC (BA9), RMPFC (BA10), RLPFC (BA10) extending to DMPFC (BA8), dorsal ACC (BA24/32). Intensity of ideation negatively correlated with goal-related activity in DMPFC (BA6/8), dorsal ACC (BA32), DLPFC (BA9), RMPFC (BA10), RLPFC (BA10), IFG (BA44/45) Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Minzenberg et al 201552 MD-P 8 SA+SI, 8 SI only 14 DC 13 (43) 18-50 Continuous performance, WB SA+SI vs SI:  activity in PCC, cuneus and precuneus and  activity right OFC (BA47), right RLPFC (BA10), premotor cortex (BA6), DLPFC (BA9/46), insula during goal- representation. Positive correlation between intensity of SI and goal-related IFG (BA45), lateral and medial OFC (BA11/47), insula and dorsal striatum activity Minzenberg et al 2015170 SCZ 8 SA+SI, 7 SI only 17 DC 6 (19) 18-50 Continuous performance, dorsal ACC seed-based FC SA+SI vs SI:  dorsal ACC FC with RMPFC and RLPFC (BA10), DMPFC (BA8), DLPFC (BA9), dorsal ACC (BA32), IFG (BA45), superior temporal gyrus, middle temporal gyrus, precuneus, and PCC during conflict monitoring. SI vs no-SI:  dorsal ACC-precuneus FC during conflict monitoring. Intensity of ideation was positively correlated with dorsal ACC FC with paracentral lobe, precuneus, left caudate, right putamen, right lateral globus pallidus and left thalamus. Minzenberg et al 2016171 MD-P 8 SA+SI, 8 SI only 14 DC 13 (43) 18-50 Continuous performance, dorsal ACC seed-based FC SA+SI vs SI:  dorsal ACC FC with left DLPFC (BA9), frontal motor areas (BA4/6), inferior temporal gyrus, middle temporal gyrus, dorsal ACC (BA24/32) during conflict monitoring. SI vs no-SI:  dorsal ACC FC with left DLPFC (BA9), DMPFC (BA8), premotor cortex (BA6), superior parietal cortex, inferior parietal cortex, superior temporal gyrus, middle temporal gyrus. Intensity of ideation positively correlated with dorsal ACC FC with bilateral premotor cortex (BA6), inferior and superior parietal cortex, middle and inferior temporal gyri, middle occipital gyrus and occipital regions, negatively correlated with IFG OFC (BA11/47), insula, putamen, globus pallidum, premotor area (BA6), somatosensory cortex (BA5/7) during conflict monitoring. Minzenberg et al 201580 SCZ 8 SA (all past SI) 9 DC (all past SI) — 18-50 Stroop, ROI frontal cortex SA vs DC:  left DPFC (BA6/8) during cognitive control. Vanyukov et al 201678 MDD 13 SA 13 DC, 22 HC 30 (63) 46-90 Delay discounting, WB SA vs DC:  left DLPFC (BA9) activation with increasing value of smaller immediate reward, with a larger decrease in people with better planned SA. Longer versus shorter delay of delayed reward associated with  left parahippocampal gyrus and middle occipital gyrus activation during trials Decision making and reward processing Pan et al 2013248 MDD 15 SA 14 DC, 13 HC 23 (55) 12-17 Iowa gambling, WB SA vs DC:  right thalamus activation during risky choices. No association between activity in thalamus and lethality of attempt and severity of SI Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Olié et al 201547 MDD 15 SA 23 DC, 35 HC 0 (0) 18-60 Iowa gambling, ROI OFC, VLPFC, MPFC, ACC, DPFC SA vs DC:  activation in left DPFC (BA8/9/46) during risky choices,  activation in bilateral DPFC, right OFC (BA11/47), right dorsal and ventral ACC (BA24/32) during winning Baek et al 201792 past MDD 10 SA 12 DC, 22 HC 24 (55) 18-44 Risk and loss aversion, ROI striatum, OFC, VMPFC, ventral ACC, midbrain SA vs DC:  subgenual ACC (BA25) activity in response to potential gain. In SA: insula activity correlated negatively with the subjective value of probabilistic gain and loss Jollant et al 2010249 MDD 13 SA 12 DC, 15 HC 0 (0) 22-59 Iowa gambling, ROI OFC, ACC, occipital cortex, precuneus/angular gyrus, thalamus, cerebellum, caudate, cuneus, superior frontal gyrus, parietal cortex SA vs DC:  lateral OFC (BA47) and occipital cortex activation during risky choices. No differences in activation during gain vs loss trials. No associations with SI Dombrovski et al 201393 MDD 15 SA 18 DC, 20 HC 31 (59) 60+ Probabilistic reversal learning, WB SA vs DC:  activity in ventral ACC (BA24/25/32) during expected reward. Poor planning of attempt associated with  activity of paralimbic network (ventral ACC/VMPFC, PCC, precuneus) during expected reward Memory Reisch et al 201072 SR-depression 8 SA 8 (100) SA: 38.5 (13.1) Recall of a mental pain, suicide action and neutral conditions using autobiographical scripts of a recent episode of SA, WB Recall of own suicidal episodes (mental pain and suicide action) vs neutral condition:  activation in left DLPFC (BA46), right RLPFC (BA10), left DMPFC (BA6),  right parahippocampal gyrus, right cuneus, left middle temporal gyrus and cerebellum. Recall of suicide action vs mental pain:  activation in the left DMPFC (BA6), right dorsal ACC (BA32) and left hippocampus Silvers et al 201650 BPD 46 SA 14 DC 60 (100) SA: 30.0 (9.8), DC: 26.7 (5.0) Reappraisal of negative autobiographical memories, ROI lateral OFC SA vs DC:  lateral OFC activation during both reappraisal and immersion, and  precuneus and cuneus activation during reappraisal of memories. In SA: regulation success associated with greater cuneus and precuneus activity Emotion processing Pan et al 201374 MDD 14 SA 15 DC, 15 HC 25 (57) 12-17 Viewing of angry, happy and neutral faces, WB activity and right dorsal ACC seed-based FC SA vs DC:  activity in right dorsal ACC, bilateral primary sensory cortex, left DLPFC (BA9), right middle temporal gyrus,  activity in insula and  FC between dorsal ACC and bilateral insula in response to angry faces. In DC: suicidal ideation was negatively correlated with left DLPFC activation during angry faces Johnston et al 201736 BD 26 SA 42 DC, 45 HC 43 (63) 14-25 Viewing of happy, neutral and fearful faces, amygdala seed- based FC SA vs DC:  FC amygdala with left OFC (BA11/47), RPFC (BA10), ventral ACC (BA32) in response to happy and neutral faces. SA lethality associated with  FC amygdala with left ventral PFC in response to happy, neutral and fearful faces. In SA: SI was negatively correlated with FC between amygdala and RPFC (BA10) Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Olié et al 201547 MDD 15 SA 23 DC, 35 HC 0 (0) 18-60 Viewing of angry, happy, sad and neutral faces, ROI OFC, VLPFC, MPFC, ACC, DPFC SA vs DC:  activation in left OFC and VLPFC in response to angry faces, and  activation in right ACC in response to sad faces Kim et al 2017136 DD 14 SA 22 HC 26 (72) 20-47 Viewing of angry, happy, sad and neutral faces, and pictures of suicidal means, WB SA vs HC:  activation in the left DLPFC, IFG, thalamus and PCC when viewing knives versus natural landscapes. No differences in activity while viewing emotional faces Jollant et al 200848 MDD 13 SA 14 DC, 16 HC 0 (0) SA: 40.3 (11.3), DC: 43.9 (10.6), HC: 32.4 (9.8) Viewing of angry, happy and neutral faces, WB SA vs DC:  activation in in right lateral OFC (BA47) and  activation in right DMPFC (BA6) in response to intense angry faces,  activation in right rostral ACC (BA32) and  activation in right cerebellum to mild happy faces,  activation in right cerebellum in response to mild angry faces Vanyukov et al 201549 MDD 18 SA 13 DC, 18 HC 26 (53) 60+ Angry and fearful faces versus shape matching, WB SA vs DC: no differences. In SA, higher activation of IFG (BA44/45) while matching angry faces was associated with poorer planning of attempt Self-referential processing Quevedo et al 2016147 DD 43 HS 39 LS, 37 HC 72 (61) HS: 14.9 (1.6), LS: 14.9 (1.8), HC: 14.5 (1.5) Emotional Self-Other Morph- Query (ESOM-Q), WB HS vs LS:  activity in RMPFC (BA10) and in a cluster including parahippocampus, hippocampus, and amygdala during happy self faces versus happy other faces. When controling for depression severity: HS vs LS:  PCC/precuneus and rostral ACC/BA10 during self faces. Social exclusion Olié et al 2017103 past MD 36 SA 41 DC, 28 HC 105 (100) 19-54 Social exclusion (Cyberball), WB SA vs DC:  activation in left supramarginal gyrus and posterior insula during social exclusion Suicidal ideation studies Resting State fMRI Cullen et al 2014250 MDD 41 DC, 29 HC 54 (77) MDD: 15.7 (2), HC: 16.0 (2) Amygdala seed-based FC No significant correlation between amygdala FC and suicidal ideation Ordaz et al 2018152 MDD 40 SI (33% past SA) 30 (75) 14-17 ICA of DMN, ECN, and SN  Network coherence of left ECN, anterior DMN and SN associated with  lifetime SI, only association with left ECN remained after controlling for depressive and anxiety symptom severity. Left ECN also associated with past SA at a trend-level Chase et al 2017199 MDD 34 SI (18 past SA) 40 HC 50 (68) 18-35 PCC seed-based FC SI vs HC:  dorsal PCC FC with left middle temporal gyrus. SA vs no-SA:  dorsal PCC FC with left IFG Du et al 2017153 MDD 28 SI 20 DC, 30 HC 55 (71) SI: 32.5 (9.9), DC: 37.1 (10.6), 35.7 (10.2) Rostral ACC seed-based FC SI vs DC:  FC right rostral ACC to medial OFC and right middle temporal gyrus, finding for right middle temporal gyrus remained when corrected for depressive symptom severity Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Kim et al 2017162 MDD 23 SI (7 past SA) 23 DC, 36 HC 68 (83) SA: 47, SI: 57, DC: 52.7, HC: 56.5 Graph theory and network- based analysis SI vs DC:  FC in a network including left lateral OFC, left and right caudate, right putamen, left middle temporal gyrus, left and right thalamus and left postcentral gyrus. Node strength, clustering coefficient, regional efficiency of lateral OFC, left and right thalamus negatively correlated with SI Cognitive control Matthews et al 201253 MDD, PTSD, TBI 13 SI 13 DC 0 (0) SI: 29.5 (4.7), DC: 27.1 (3.6) Stop signal response inhibition, WB SI vs DC:  left dorsal ACC (BA24), RLPFC (BA10), DLPFC (BA9/46), supramarginal gyrus, OFC (BA11), DMPFC (BA6) and superior temporal gyrus during error processing Lee et al 201581 SCZ 28 DC, 17 HC 8 (18) SH: 43.6 (11.3), DC: 38.9 (7.3), HC: 37.9 (12.9) Go-no-go response inhibition, WB Positive correlation right DLPFC (BA9) activity during response inhibition and current SI in self-harm but no association in no-self-harm group Zhang et al 2013251 SCZ 14 SI 19 DC, 15 HC 24 (50) 18-45 N-back, dynamic causal modelling with PCC and MPFC seeds High vs Low risk:  no difference in activation of and connectivity between VMPFC and PCC. VMPFC activity was positively related with suicide risk Decision making and reward processing Quevedo et al 2017161 MDD 38 DC, 30 HC 44 (65) DC: 30.7 (7.7), HC: 32.0 (6.1) Card guessing, ventral striatum seed-based FC Positive correlation between SI and left ventral striatum FC with DMPFC, DLPFC and dorsal ACC during loss trials. Motor control Marchand et al 2012138 MDD 5 SA 17 DC 0 (0) 22-45 Motor activation, putamen seed- based FC Positive correlation between SI and left putamen FC with left DMPFC and right putamen, and right putamen FC with left putamen. Left putamen FC with DMPFC also associated with depressive symptom severity Marchand et al 2013252 MDD, BD 22 SI 18 DC ND 21-45 Motor activation, PCC seed- based FC Positive correlation between SI and PCC FC with left DLPFC, DMPFC and IFGs in MDD but not BD. PCC FC with IFG and DLPFC also associated with depressive symptom severity Marchand et al 2011135 BD-II 10 SI 6 DC, 19 HC 0 (0) BD-II: 32.9 (7.5), HC: 33.7 (12.5) Motor activation, ROI putamen Negative correlations between a history of SI and activation in left putamen Emotion processing Marchand et al 2011117 BD 10 past SI (3 past SA) 6 DC, 19 HC 0 (0) 21-60 Viewing of happy, fearful and neutral faces, ROI amygdala and subgenual ACC No correlation between brain activity during the task and history of SI Authors, year Mental disorder SA group/ SI group Groups w/o SA and/or SI Female n (%)* Age Methods Findings** Just et al 201860 No DX 17 SI (9 past SA) 17 HC 26 (77) SI: 22.9 (3.6), HC: 22.1 (2.8) Neurosemantic analyses of concepts related to suicide, positive and negative affect, machine learning on voxels with stable semantic tuning curves SI could be discriminated from HC with 91% accuracy. Most discriminating regions were the left VMPFC, left DMPFC extending to dorsal ACC, right middle temporal gyrus, left inferior parietal cortex, and left IFG. SI+SA group could be discriminated from SI without SA with 94% accuracy, with most discriminating regions including the left VMPFC, left DMPFC extending to dorsal ACC, right middle temporal gyrus. Emotion regulation Miller et al 201873 No DX 14 SI (4 past SA) 32 without SI 29 (63) 13-20 Emotion regulation, WB SI vs no-SI:  activity in thalamus, IFG/DLPFC (BA44/9), temporoparietal junction and cerebellum and  activity in temporal pole during passive viewing of negative pictures,  activity in DLPFC (BA9) during regulation of negative emotional pictures NEAR-INFRARED SPECTROSCOPY STUDIES Suicide attempt studies Tsujii et al 2017253 MDD 30 SA 38 DC, 40 HC 69 (64) SA: 37.6 (10.0), DC: 38.8 (9.7), HC: 38.2 (10.2) Verbal fluency SA vs DC:  verbal fluency task induced changes in mean oxy-HB in left precentral gyrus Suicidal ideation studies Pu et al 2015254 MDD 31 SI 36 DC, 67 HC 76 (57) SI: 57.3 (15.7), DC: 58.7 (16.5), HC: 58.1 (17.8) Verbal fluency SI vs DC:  verbal Fluency task induced changes in oxy-Hb in right DLPFC, lateral OFC and right RLPFC Symbols & Abbreviations: *Percentages are rounded to the nearest whole number; **Results are reported for SA or SI in comparison with diagnostic controls. If no diagnostic controls were included in the study, results based on SA or SI compared to healthy controls are reported; ACC: anterior cingulate cortex; BA: Broadman’s Area; BD: bipolar disorder; BD-II: bipolar II disorder; BPD: borderline personality disorder; DC: diagnostic controls; DD: depressive disorder; DLPFC: dorsolateral prefrontal cortex; DMN: default mode network; DMPFC: dorsomedial prefrontal cortex; ECN: executive control network; FC: functional connectivity; HC: healthy controls; HS: high suicidality; ICA: independent component analysis; IFG: inferior frontal gyrus; LS: low suicidality; MDD: major depressive disorder, MD-P: psychotic mood disorder, MPFC: medial prefrontal cortex; ND: not detailed; OFC: orbitofrontal cortex; oxy-HB: oxygen-hemoglobin; PCC: posterior cingulate cortex; PFC: prefrontal cortex; PTSD: posttraumatic stress disorder; ReHo: regional homogeneity; RLPFC: rostrolateral prefrontal cortex; RMPFC: rostromedial prefrontal cortex; SA: suicide attempt; SCZ: schizophrenia; SI: suicidal ideation; SN: salience network; SR-depression: self-reported depression; ROI: region of interest; TBI: traumatic brain injury; VLPFC: ventrolateral prefrontal cortex, VMPFC: ventromedial prefrontal cortex; WB: whole brain; zALFF: z score amplitude of low frequency fluctuations