There are many scientists in the field of suicidology who believe that there are reliable differences in the way the brain functions in people who think about or attempt suicide compared to those who don’t [1-3]. This belief that suicidal thoughts and behaviors, and psychiatric conditions in general, are driven by neural aberrations is reflected in the National Institute of Mental Health’s (NIMH) funding priorities in recent years. For example, NIMH unveiled the Research Domain Criteria framework, which assumes that psychiatric disorders are “brain disorders” at the brain circuit level . A key question though is what the literature says about suicidality and the brain. That is, are there really brain differences in those who experience suicidality? To discuss this, I’ll focus on two of the main ways we examine how the brain is functioning in humans: fMRI and event-related potentials.
fMRI and Suicidality
fMRI is the method most people are familiar with when discussing neuroscience research in humans. Briefly, fMRI is a way of indirectly measuring brain activity by imaging the hemodynamic response. If a brain area is working harder, it needs more oxygen, more blood will flow to that area, and this flow is quantified to infer when different areas are increasing or decreasing in neural activity during various tasks. Two reviews have recently examined the research using fMRI to determine whether there are brain differences in those experiencing suicidality. The first was a narrative review, and they determined that there are two networks associated with suicidal thoughts and behaviors . First, the authors proposed that the ventral prefrontal cortex and many of its connections are involved in increasing negative and decreasing positive internal states, which may lead to suicidal ideation. Second, they identified other regions, including the dorsomedial prefrontal cortex, the dorsolateral prefrontal cortex, and the inferior frontal gyrus, as a separate network that may be important for suicide attempts due to these regions’ roles in planning and cognitive control. A second review conducted a meta-analysis rather than a narrative review, and only included studies that used whole-brain analyses . They did not find significant associations between suicide attempts and any brain region. They also concluded that there were not enough studies to analyze the relationship between suicidal ideation and neural function. Based on these two reviews, it is unclear whether there are neural differences in those experiencing suicidality.
ERPs and Suicidality
Event-related potentials (ERPs) are derived from electroencephalographic (EEG) data and reflect neural responses to specific, repeated events . ERP waveforms are calculated by averaging the electrocortical response to a specific event (e.g., stimulus presentation, error commission) across multiple trials. By averaging the EEG waveform before, during, or after a specific event that is repeated, the noise inherent in EEG is minimized, creating smoother waveforms that reflect the electrocortical activity in the brain. Some colleagues and I recently conducted a meta-analysis to examine whether there are neural differences in those experiencing suicidality as measured by ERPs . Depending on the outcome analyzed (e.g., suicidal ideation, previous suicide attempt, or suicide risk), we found small-to-moderate effects. Based on this information alone, you may be tempted to conclude that there are neural differences in those experiencing suicidality if you measure neural activity using ERPs. The reality, however, is much more complicated.
Statistical power is the probability that you will find an effect (i.e., the effect was “statistically significant”) of a given size if there is truly an effect of that size (or larger) in the population. Some ways to increase power are to increase sample size and by minimizing error/noise in measures. By convention, a study is considered to be at least acceptably powered when power is 80%. However, research shows that insufficient power is the norm in fMRI and ERP studies. Specifically, most fMRI studies have less than 50% power, and most ERP studies are only sufficiently powered to detect large effects which are considered unlikely for most research questions [9,10]. When studies are insufficiently powered, this not only increases the probability of the study finding a false negative (i.e., the study does not find an effect when there is a real effect in the population), it also increases the probability of false positives (i.e., the study says there an effect, when there is actually no effect) . fMRI and ERP studies examining suicidality appear to be no exception, with the median sample size for fMRI studies being 45 participants, and the average for ERP studies being 73 participants [6,8].
Another barrier to answering whether there are neural differences in those experiencing suicidality is the low reliability of fMRI measures. This was shown by a recent meta-analysis which found that the test-retest reliability of most fMRI measures were poor (i.e., .397), which is far below the minimum cutoff for good reliability, not to mention clinical application or individual interpretation . Because suicidology is by-in-large a field that studies individual differences (i.e., correlational rather than experimentally-induced relationships), this low reliability is problematic. This is because reliability is critical to detect between-subjects effects . Why are fMRI measures this way? There are many reasons, but one is that most tasks were designed to minimize individual differences and maximize within-subjects differences. Thus, while robust within-subjects task effects are ample throughout cognitive neuroscience, these robust within-subjects effects do not necessarily translate to reliable between-subjects effects . Though ERPs suffer from a similar problem, it appears to not be as large of an issue, with many ERPs demonstrating good internal consistency [15,16].
So… What Does it Mean?
Based on the studies I have discussed, I conclude that there is no strong evidence that suicidality is related to neural aberrations. But there is also no strong evidence that suicidality is not related to neural aberrations. In other words, we don’t know. Most studies are too underpowered to say one way or the other, and tasks have been designed to maximize within-subjects differences at the cost of being able to detect individual differences. To address these issues, there needs to be a fundamental shift in the priorities of this field. First, neuroscientists should focus on creating consistent definitions for suicidal outcomes and constructs to be measured by fMRI and ERPs. Second, the field should develop tasks with strong psychometric properties that are designed to detect individual differences. Third, researchers need to collaborate on a large scale to achieve sufficient statistical power in their studies. At least one consortium already exists to use fMRI to study suicidal thoughts and behaviors, but large datasets alone will not make the research meaningful. Without first paying greater attention to the psychometric properties of their measures, the public may begin to question what the neuroscience of suicidal thoughts and behaviors has to offer.
 Joiner, T. E., Brown, J. S., & Wingate, L. R. (2005). The psychology and neurobiology of suicidal behavior. Annual Review of Psychology, 56, 287–314. https://doi.org/10.1146/annurev.psych.56.091103.070320
 Mann, J. J. (2003). Neurobiology of suicidal behaviour. Nature Reviews: Neuroscience, 4, 819–828. https://doi.org/10.1038/nrn1220
 Van Heeringen, K., & Mann, J. J. (2014). The neurobiology of suicide. The Lancet Psychiatry, 1(1), 63–72. https://doi.org/10.1016/S2215-0366(14)70220-2
 Insel, T., Cuthbert, B., Garvey, M., Heinssen, R., Pine, D. S., Quinn, K., Sanislow, C., & Wang, P. (2010). Research domain criteria (RDoC): Toward a new classification framework for research on mental disorders. American Journal of Psychiatry, 167(7), 748–751.
 Schmaal, L., van Harmelen, A.-L., Chatzi, V., Lippard, E. T. C., Toenders, Y. J., Averill, L. A., Mazure, C. M., & Blumberg, H. P. (2019). Imaging suicidal thoughts and behaviors: A comprehensive review of 2 decades of neuroimaging studies. Molecular Psychiatry. https://doi.org/10.1038/s41380-019-0587-x
 Huang, X., Rootes-Murdy, K., Bastidas, D. M., Nee, D. E., & Franklin, J. C. (2020). Brain Differences Associated with Self-Injurious Thoughts and Behaviors: A Meta-Analysis of Neuroimaging Studies. Scientific Reports, 10(1), 2404. https://doi.org/10.1038/s41598-020-59490-6
 Luck, S. J. (2014). An introduction to the event-related potential technique (2nd ed.). The MIT Press.
 Gallyer, A. J., Dougherty, S. P., Burani, K., Albanese, B. J., Joiner, T. E., & Hajcak, G. (2020). Suicidal thoughts, behaviors, and event-related potentials: A meta-analysis. BioRxiv. https://doi.org/10.1101/2020.04.29.069005
 Clayson, P. E., Carbine, K. A., Baldwin, S. A., & Larson, M. J. (2019). Methodological reporting behavior, sample sizes, and statistical power in studies of event‐related potentials: Barriers to reproducibility and replicability. Psychophysiology. https://doi.org/10.1111/psyp.13437
 Nord, C. L., Valton, V., Wood, J., & Roiser, J. P. (2017). Power-up: A Reanalysis of “Power Failure” in Neuroscience Using Mixture Modeling. The Journal of Neuroscience, 37(34), 8051–8061. https://doi.org/10.1523/JNEUROSCI.3592-16.2017
 Button, K. S., Ioannidis, J. P. A., Mokrysz, C., Nosek, B. A., Flint, J., Robinson, E. S. J., & Munafò, M. R. (2013). Power failure: Why small sample size undermines the reliability of neuroscience. Nature Reviews Neuroscience, 14. https://doi.org/10.1038/nrn3475
 Elliott, M. L., Knodt, A. R., Ireland, D., Morris, M. L., Poulton, R., Ramrakha, S., Sison, M. L., Moffitt, T. E., Caspi, A., & Hariri, A. R. (2020). What Is the Test-Retest Reliability of Common Task-Functional MRI Measures? New Empirical Evidence and a Meta-Analysis. Psychological Science, 31(7), 792–806. https://doi.org/10.1177/0956797620916786
 Hajcak, G., Meyer, A., & Kotov, R. (2017). Psychometrics and the neuroscience of individual differences: Internal consistency limits between-subjects effects. Journal of Abnormal Psychology, 126(6), 823–834. https://doi.org/10.1037/abn0000274
 Infantolino, Z. P., Luking, K. R., Sauder, C. L., Curtin, J. J., & Hajcak, G. (2018). Robust is not necessarily reliable: From within-subjects fMRI contrasts to between-subjects comparisons. NeuroImage. https://doi.org/10.1016/j.neuroimage.2018.02.024
 Meyer, A., Riesel, A., & Proudfit, G. H. (2013). Reliability of the ERN across multiple tasks as a function of increasing errors: Reliability of the ERN across multiple tasks. Psychophysiology, 50(12), 1220–1225. https://doi.org/10.1111/psyp.12132
 Moran, T. P., Jendrusina, A. A., & Moser, J. S. (2013). The psychometric properties of the late positive potential during emotion processing and regulation. Brain Research, 1516, 66–75. https://doi.org/10.1016/j.brainres.2013.04.018