The Effect of Childhood Abuse on Cognitive Performance in patients with Bipolar Disorder

Rachel Van Boxtel, BS

Brigham and Women’s Hospital
The Effect of Childhood Abuse on Cognitive Performance in patients with Bipolar Disorder

Scientific Abstract

Background: Early-life trauma is prevalent in Bipolar Disorder (BD) and plays a major role in the course of the disease. Early-life stressors can adversely affect regions in the developing cognitive brain, particularly hippocampal circuity. Another common feature of bipolar disorder is cognitive deficits, namely in executive function, working memory, attention, and processing speed. Prior work has indicated that childhood trauma is associated with neurocognitive impairment in BD patients, but with modest sample sizes. In this study, we further investigated the impact of childhood trauma on performance on neurocognitive measures.

Methods: 259 Adults with DSM-IV BD-I and II were recruited from the Icahn School of Medicine at Mount Sinai Hospital and Brigham and Women’s Hospital. Diagnostic eligibility was confirmed using The Structured Clinical Interview for the DSM-V (SCID-V). Cognitive performance was assessed using a battery of neurocognitive assessments. Statistical measures included a multivariate analysis of covariance to assess strength and significance of associations between childhood abuse and performance on neurocognitive measures (controlled for level of education).

Results: A sizable portion of our participants with BD (N=259) reported emotional abuse (43%), physical abuse (31%), sexual abuse (35%), emotional neglect (33%), and physical neglect (45%) on the Childhood Trauma Questionnaire (CTQ). A multivariate analysis of covariance revealed significant effects of physical abuse (PA) (F (11, 207) =2.3, p=0.01) and emotional abuse (EA) (F (11, 207) =2.1, p=0.02) on cognitive performance across neurocognitive domains. PA was negatively associated with performance on Social Cognition (F=7.0, p=0.009), the Stroop (F=5.5, p=0.02), and Reading Mind in the Eyes (F=5.0, p=0.03). EA was positively correlated with performance on Social Cognition (F=6.4, p=0.01).

Conclusions: BD patients with self-reported “severe” childhood abuse (PA and EA) have marked changes in neurocognitive performance, particularly in Social Cognition where the largest effect sizes were observed.

SoundCloud Transcript

Hi, my name is Rachel Van Boxtel, and I’m going to be walking you through my poster today on the effect of childhood abuse on cognitive performance in patients with bipolar disorder. As some background to the topic, early-life trauma is very prevalent in bipolar disorder and can play a major role in the course of illness. Stress such as this can impact cognitive regions of the developing brain, particularly in hippocampal circuitry. And cognitive deficits are another hallmark feature of bipolar disorder, particularly seen in executive functioning and working memory. Prior work has shown the effect of childhood abuse and its association with neurocognitive impairment in adults with bipolar disorder, but with modest sample sizes. So, our goal was—with a larger sample size—to examine whether a history of childhood abuse affects various aspects of neurocognitive performance in patients with bipolar disorder.

259 adults with bipolar disorder were part of our study—with bipolar I and bipolar II diagnoses—and we confirmed their diagnostic eligibility using the SCID. These participants were given a battery of neurocognitive assessments, including the MATRICS battery, the Stroop, the Controlled Oral Word Associated Task, and Reading Mind in the Eyes task. Participants also completed the Childhood Trauma Questionnaire. And for the analyses of these data, we used multivariate analyses of covariance in order to assess for both strength and significance of associations between varying forms of childhood abuse and performance across neurocognitive measures. These analyses were controlled for education, and then further analyses were done to determine the directionality of these effects.


So, going into the results here, abuse was prevalent within our sample with 31% of participants reporting physical abuse, 43% reporting emotional abuse, 35% sexual abuse, 33% emotional neglect, and 45% physical neglect. For the results of the multivariate tests, both physical abuse and emotional abuse history were associated with cognitive performance across domains of the neurocognitive assessments. So, breaking it down into the neurocognitive assessments themselves, for physical abuse, there was a significant negative association between performance on the Stroop, Reading Mind in the Eyes, Reasoning and Problem Solving from the MATRICS battery, and also Social Cognition. And for emotional abuse, there was a positive correlation with Social Cognition compared to healthy controls. Um, but then we decided from there to control for further covariates besides for just education. So, we also included sex, age, race, and current mood severity determined by the HAM-D and the Young Mania Rating Scale, and physical abuse remained negatively associated with performance on the Stroop, the MCCB reasoning and problem solving, and Reading Mind in the Eyes. Um, however, emotional abuse was no longer significant as this result broke down. So as a summary of our results, a sizeable portion of our participants with bipolar disorder reported a history of abuse on the CTQ. Multivariate analyses revealed significant associations, um, between physical abuse and emotional abuse history on neurocognitive performance across domains. Physical abuse history was associated with a variety of measures, and emotional abuse history was positively correlated with Social Cognition, but again, this result was no longer significant when considering other covariates. So, for thinking more broadly, it’s clear from our data that patients with self-reported severe childhood abuse have marked differences in their cognitive performance – particularly with physical abuse where these effects where well-controlled across a variety of demographic covariates as well as, um, current mood severity. And then, finally, we can think about this and how it lends itself towards future implications for improving treatment and a continued importance of trauma-informed care for patients with bipolar disorder. Thank you for your time.

Live Zoom Session – April 21st

research Areas


Rachel Van Boxtel, BS, Caitlin Millett, PhD, Katherine Burdick, PhD

Principal Investigator

Katherine Burdick, PhD