The Impact of Intravenous Ketamine on Attentional Bias: Probing Mechanisms of Rapid-Acting Antidepressant Effects Across Two Clinical Studies

Abstract

Background:

Ketamine is known for its rapid antidepressant effect, but its impact on affective information processing (including attentional bias, a putative cognitive mechanism of depression), remains largely unexplored. We leveraged a novel measurement of attentional bias and sought to: (1) establish adequate test-retest reliability and validity among depressed participants prior to ketamine treatment; and (2) harness a single dose of ketamine to assess mechanistic shifts in attentional bias and their relation to antidepressant efficacy.

Methods:

A novel dual probe video task was used to index attentional bias toward sad film clips. In Study 1, treatment-seeking adults with moderate-to-severe depression (n=40) completed the task at 1) baseline, 2) 1-week retest, 3) 1-month retest, and, for a subset (n=15), 4) 24-hrs post-ketamine infusion (0.5mg/kg over 40min). In Study 2, participants (n=43) completed the task pre- and 24-hrs post-ketamine.

Results:

Indices from the novel attentional bias task were stable prior to ketamine, demonstrating good one-week and one-month test-retest reliability. Participants in both studies exhibited a robust reduction in attentional bias from pre- to 24-hrs post-ketamine infusion. In Study 1, cross-sectional correlations were observed between attentional bias and clinician-rated depressive symptoms at each pre-treatment assessment. In Study 2, changes in attentional bias were correlated with improved symptoms from pre- to post-infusion.

Conclusions:

Results provide evidence for the validity of a novel, psychometrically robust measure of attentional bias among individuals with depression. Findings indicate that ketamine reliably and rapidly reduces attentional bias, offering insight into a replicable, potential cognitive mechanism involved in its antidepressant action.

Study 1: clinicaltrials.gov/study/NCT04916548; Study 2: clinicaltrials.gov/study/NCT05168735

Introduction

Lifetime prevalence rates of depressive disorders affect ~20% of the U.S. population(1) and depression is a leading contributor to the global burden of disease(2). While efficacious treatments are available, remission rates remain low even following evidence-based treatment(3). Patients who do not respond to current frontline therapy are at increased risk for chronic symptoms, suicidality, disability, and a significantly shorter life expectancy(4). To spur treatment innovation, future research is imperative to gain a deeper understanding of the mechanisms underlying depression and treatment response.

An emerging class of rapid-acting antidepressants, including intravenous ketamine, produce swift and robust antidepressant effects(5), even among patients with challenging-to-treat psychiatric disorders such as treatment-resistant depression(6). Pharmacotherapies such as ketamine also represent a powerful experimental tool that can provide a more precise evaluation of neurocognitive mechanisms underlying reductions in clinical depressive symptoms(7). In randomized controlled trials, intravenous ketamine at subanesthetic doses (typically, 0.5mg/kg) consistently produces a reliable, fast-acting, and potent antidepressant effect (metanalytic Cohen’s d=1.4)(8). The antidepressant effects of ketamine begin approximately 2 hours post-infusion, following the end of acute dissociative and euphoric side effects. These effects persist well beyond the drug’s elimination half-life of 2.5-3 hours, typically enduring for 3-14 days(9-11) .

The mechanisms driving ketamine’s antidepressant effects remain largely unknown(7), and uncovering factors that drive its impact may spur the development of novel adjunctive treatments that can extend its antidepressant effects(12). To that end, there is considerable interest in using ketamine to experimentally induce changes in depressive symptoms and subsequent neurocognitive functioning(7), but to date, the vast majority of studies have focused on one of two extremes: molecular/neuronal levels of analysis (e.g., using animal models of depression) or clinical efficacy. Relatively few studies have examined ketamine’s impact on affective information processing patterns, which are one key factor situated along a posited integrative continuum that bridges from molecules to patient-reported mood states(7)—and virtually no reported effects on these neurocognitive measures have been replicated across more than one independent patient sample.

Cognitive theorists of depression have long posited that systematic alterations in selective attention to affective information (i.e., attentional bias) is a key neurocognitive contributor to chronically depressed mood(13). The relation between depression and attentional bias is critical in part because attention provides a selective filter of the source material from our internal and external environments that will remain available for further neurocognitive processing(14). Specifically, individuals with depression tend to exhibit a bias for allocating attention toward negative environmental information while diverting it from benign information(15). This putatively leads to a skewed integration of information across neurocognitive systems, resulting in a stream of predominantly negative recollections and interpretations of the self, others, and the world, which in turn creates a vicious feedback loop that reinforces sustained focus on negative information and persistence of depressed mood.

To assess the patterns of attentional bias that characterize depression, most studies have employed the “dot-probe” task(16), during which participants respond to the location of a probe appearing behind either the location of an emotional or benign stimulus that were displayed simultaneously. Biased attention toward emotional stimuli, indicated by faster reaction times when the probe appears in the emotional stimulus location, suggests preferential attention for those stimuli compared to benign ones. Case-control studies using the dot probe task have demonstrated that, compared to individuals with no depression history, those with depression show preferential attention for sad faces(e.g., 17) and negative, depressogenic words(e.g., 18), with medium meta-analytic effect sizes[d=.52;(19)]. Although these are well-replicated findings, the field has grappled with acknowledged limitations of the task(20, 21), including: a) low psychometric reliability and b) reliance on static stimuli (words, faces) that may lack ecological validity. These significant concerns have raised questions about replicability and have stymied the clinical implications that can be drawn from this body of work.

To address the described limitations of the dot-probe task, alternative assessments have been developed to quantify attentional bias with greater reliability and ecological validity. Specifically, Grafton, Teng, and MacLeod (2021) recently introduced an innovative dual probe task to assess attentional bias in anxiety, which may offer similar benefits for depression research. Instead of presenting a single probe on each trial in the location of either the negative or benign member of a stimulus pair, the dual probe method simultaneously presents two probes, one positioned in the location of the negative stimulus, another in the location of the benign stimulus. The proportion of correctly identified probes in the location of negative stimuli offers an indicator of negative attentional bias, with high psychometric reliability (internal consistency=.97)(22). Grafton and colleagues also demonstrated that when dual probes are displayed, the task can be used to effectively detect attentional biases associated with anxiety. Because the dual probe approach is compatible with continuous video stimuli, it also allows for the presentation of more ecologically-valid negative content.

To address the longstanding limitations in mechanistic investigations of attentional biases, we employed a novel variant of the dual probe task in two separate studies. Treatment-seeking adults with moderate-to-severe depression completed the novel task where they were asked to watch competing video clips from commercially available movies, focused on either sad or benign emotional scenes. In Study 1, they completed the task four times – 1) initial assessment, 2) 1-week follow-up, 3) 1-month follow-up, and, 4) for a subset of participants (n=15), 24-hrs post-ketamine infusion. In Study 2, they completed the task pre- and 24-hrs post-infusion. Two prior studies, to our knowledge, have used a similar pre- and post-ketamine design to assess changes in behavioral indices of attentional bias using the dot probe task. Both studies yielded null results and did not report on key psychometric properties such as test-retest reliability(23, 24). In contrast, we proposed to explicitly examine whether there was acceptable test-retest reliability and convergent validity for the attentional bias measures extracted from our novel dual probe video task, as assessed across assessments delivered over a ~1 month interval with no intervening changes to treatment (Study 1). Assuming acceptable psychometric properties - an essential pre-condition for repeated measurements of intervention outcome - we hypothesized that attentional bias would be significantly ameliorated by a single dose of ketamine and that the magnitude of improvement in attentional bias would be associated with the magnitude of ketamine’s antidepressant effects (replicable findings across Studies 1 & 2).

Method

Participants

For Study 1 (n=40; clinicaltrials.gov:NCT04916548) and Study 2 (n=43; NCT05168735), we recruited adults ages 18 to 60 years with clinically elevated, unipolar depressive symptoms [Hamilton Depression Rating Scale score ≥14(25)]. All enrolled participants in both studies met DSM-5 diagnostic criteria for Major Depressive Disorder (94%) or Depressive Disorder Not Otherwise Specified (6%), per the MINI diagnostic interview administered at baseline by an interviewer trained to reliability (ICC≥.9), with final diagnostic determinations made by a licensed clinical psychologist (RBP). See Supplement for exclusion criteria and CONSORT diagrams. Table 1 presents demographic and clinical characteristics for each study. Participants in Study 2 were significantly older and more likely to be using psychotropic medications than participants in Study 1. Patient differences in medication use were due to differential exclusion criteria that disallowed psychotropic medication for participants included in Study 1’s ketamine phase subsample but not in Study 2.

Table 1.

Demographic and Clinical Characteristics

	Study 1	Study 2	t/χ2
Age	32.13 (10.60)	38.07 (13.51)	−2.18*
Current Use of Psychotropics	47%	88%	15.90**
DSM-5 depression diagnosis			0.14
Major Depressive Disorder	95.1%	93.2%
Depressive Disorder NOS	4.9%	6.8%
Median Household Income	$50,000	$44,000	.19
Female Assigned at Birth	55%	65%	.82
Asian American	11%	2%	2.34
Black/African American	14%	7%	.87
Multiracial	3%	12%	2.38
White	73%	79%	.70
Hispanic/Latine	11%	10%	.04
Initial Assessment AB	.54 (.19)	-	-
1-Week Retest AB	.54 (.19)	-	-
1-Month Retest/Pre-Infusion AB	.55 (.20)	.59 (.18)	−.90
24-hrs Post-Infusion AB	.41 (.17)	.42 (.19)	−.27
Initial Assessment MADRS	29.87 (6.39)	-	-
1-Week Retest MADRS	27.97 (7.09)	-	-
1-Month Retest/Pre-Infusion MADRS	28.30 (7.55)	28.81 (7.28)	−.31
24-hrs Post-Infusion MADRS	11.80 (7.55)	12.90 (8.33)	−.45

Note. To ascertain race and ethnicity, participants were asked to self-report these identities. Among the four individuals who identified as Hispanic/Latine in Study 1, 50% self-identified their race as Black, 25% as White, and 25% chose not to disclose; Among the four individuals who identified as Hispanic/Latine in Study 2, 75% self-identified their race as Multiracial and 25% as White.

To ascertain sex and gender, participants were asked to self-report these identities. Among the 17 individuals who identified that they were assigned male at birth in Study 1, 94% self-identified their gender as male and 6% declined to report. Among the 21 individuals who identified that they were assigned female at birth in Study 1, 100% self-identified their gender as female. Among the 15 individuals who identified that they were assigned male at birth in Study 2, 93% self-identified their gender as male and 7% as female. Among the 28 individuals who identified that they were assigned female at birth in Study 2, 100% self-identified their gender as female.

AB = Attentional Bias (scores greater than .50 indicate a bias for sad videos); MADRS = Montgomery-Åsberg Depression Rating Scale.

^*p < .05

^**p < .001

Despite Study 1 and 2 being larger parent trials each with their own primary aims, which will be described in forthcoming manuscripts, due to the uniformity of numerous methodological features and the contemporaneous completion of the trials, the current analysis provided a unique opportunity to assess replicability and reproducibility in ketamine’s impact on affective information processing. Specifically, due to our enduring interest in examining ketamine’s influence on attention bias(7), compiling data from both studies (in this—the first empirical report from either study) allowed us to rigorously evaluate the psychometrics and clinical malleability of a novel measure of attention bias, which served as a uniform, pre-specified mechanistic outcome measure included in both studies.

Assessment Timeline

In Study 1, participants completed 3 waves of naturalistic assessments that included administration of the dual probe task [initial assessment; 1-week retest; 1-month retest]. In an open-label design, a subsample (n=15; see Supplement) received a single 40-minute infusion of intravenous ketamine (0.5mg/kg) and completed the dual probe task at both pre-infusion (the 1-month retest datapoint was used as the pre-infusion measure and was uniformly collected within 1 week of infusion) and 24-hrs post-infusion.

In Study 2, 43 participants received a single 40-minute infusion of ketamine (0.5mg/kg). Though not our primary focus given the lack of anticipated relevance to attentional bias per se, participants in the parent trial were randomized to one of two infusion contexts: a) usual/typical infusion conditions (identical to those used consistently in Study 1) or b) adjunctive mindfulness training (see Supplement for rationale for the randomized controlled trial component and mindfulness procedures). Participants completed the dual probe task within 1 week prior to ketamine infusion and 24-hours post-infusion.

Measures

Clinician-Rated Depressive Symptoms.

Each time the dual probe task was completed in both studies, a clinical rater trained to reliability (individual item inter-rater ICC’s≥.9) administered the Montgomery-Åsberg Depression Rating Scale (MADRS)(26). An additional pre-infusion MADRS score was collected on infusion morning, preceding intervention procedures. To quantify individual changes in depression from pre- to post-infusion, a percent change score was calculated as: (pre-infusion – post-infusion) / pre-infusion MADRS total score, with scores greater than 0 reflecting improvement in depression from pre-infusion baseline.

Dual Probe Video Task.

The design of the novel dual probe task was consistent with previous research(22). The task included 12 dual video pairs, each comprising one sad and one benign clip sourced from commercially available movies. Each 90-second pair was positioned to the left and right of the screen center, with intermittent switches in position every 5-7 seconds. Two visual probes, displayed for 200 ms after each switch, enabled participants to identify the probe corresponding to the attended video. Despite instructions informing participants of identical probe positions on both sides, the location of probes for sad vs. benign clips was unique. There were 16 probe pairs displayed during each of the 12 dual video pairs (192 total trials). Additional task details can be found in the Supplement.

Attentional bias (AB) for sad videos was indexed by calculating the proportion of correctly identified probes that interrupted the sad video clip of the dual video pairs. Consequently, we calculated AB for sad videos as: Number of correctly identified probes interrupting sad video clips/Total number of correctly discriminated probes. A higher score indicates a stronger AB.

To calculate percent change in AB from pre- to post-infusion (in parallel to MADRS calculations), we subtracted post-infusion from pre-infusion scores and then divided this difference by the pre-infusion score. Participants who exhibited a percent change greater than zero were categorized as displaying an ‘improvement’ (i.e., a decrease) in AB following infusion.

After excluding data from assessments with low accuracy or a strong ‘position bias’ (i.e., almost exclusively looking at one side of the screen regardless of video content)(see Supplement for details), in Study 1, 30 participants’ data remained for subsequent analysis at initial assessment, 32 at 1-week retest, 33 at 1-month retest, and 14 at 24-hrs post-infusion. In Study 2, 27 participants’ data remained at pre-ketamine infusion and 29 at 24-hrs post-infusion.

Missing Data Analysis.

Data were found to be missing completely at random (MCAR) in both Study 1, χ²=49.111, p=.626, and Study 2, χ²=17.250, p=.370. Missing data were handled using pairwise deletion in all subsequent analyses. This approach allowed the use of all observed data and was chosen because using estimated values may have generated unintended bias, given that attentional bias assessments with low accuracy or a strong position bias were systematically excluded(27).

Analytic Plan.

An inspection of our data revealed that there was no significant skew for any study variable [z>2.5;(28)]. In Study 1, to determine test-retest reliability, we calculated intraclass correlation coefficients across the 3 waves of naturalistic assessments, and Pearson correlations were used to test the cross-sectional associations between MADRS and AB scores as a measure of convergent validity during these assessments. In both Study 1 and 2, ketamine effects were explored using paired sample t-tests to examine pre- to post-infusion changes in MADRS and AB, while associations between ketamine-induced changes in MADRS and AB were tested using Pearson correlations.

Results

Table 1 presents the mean (SD) for MADRS and AB scores at each study time point.

Study 1: Psychometric Analysis.

Test-Retest Reliability.

In Study 1, test-retest correlations of AB scores were significant from initial assessment to 1-week retest (ICC=.717, p<.001) and from 1-week to 1-month retest (ICC=.753, p<.001). Similar effects were found when examined across the initial, 1-week retest, and 1-month retest assessments simultaneously (ICC=.684, p<.001). These results indicate adequate test-retest reliability in the ‘good’ range by conventional psychometric standards(29, 30).

Convergent Validity.

As seen in Figure 1, in Study 1 AB scores were correlated with MADRS scores in the expected direction at the initial assessment (r=.397, p=.030), 1-week retest (r=.389, p=.028), and the 1-month retest (r=.387, p=.026).

Figure 1. Cross-Sectional Correlations Between Attentional Bias and MADRS Scores during Study 1 Test-Retest Phase.

Studies 1 and 2: Ketamine Effects.

Pre-Post Ketamine Effect on Depressive Symptoms.

Clinician-rated depressive symptoms decreased from pre-infusion to 24-hrs post-infusion in Study 1, t(14)=7.216, p<.001, Cohen’s d=1.863, mean % change=63% (SD=23%), and Study 2, t(38)=10.469, p<.001, Cohen’s d=1.643, mean % change=58% (SD=25%).

Pre-Post Ketamine Effect on Attentional Bias.

In Study 1, AB decreased from pre-infusion to 24-hrs post-infusion, t(13)=2.538, p=.025, Cohen’s d=.678. These findings were replicated in Study 2, as AB decreased from pre- to 24-hrs post-infusion, t(21)=4.789, p<.001, Cohen’s d=1.021. Across both studies, 89% of participants (32/36) displayed a decrease in AB from pre- to post-infusion, measured over approximately one week. In contrast, across Study 1’s initial assessment to 1-week retest (same interval of ~1 week), only 56% of participants (15/27) exhibited a decrease. As visually depicted in Figure 2, in an observational (non-randomized) post hoc analysis, the type of 1-week interval (initial assessment/no-ketamine vs. pre-to-post-ketamine) was associated with the likelihood of experiencing a decrease in AB across two adjacent assessments (Chi-squared=9.05, p=.007).

Figure 2. Change in Attentional Bias Score, as a Function of Interval.

Note. AB = Attentional Bias. In this graph, interval is defined as either no-intervention or pre-post ketamine infusion 1-week interval. For visualization purposes, participants’ change score is plotted against the zero point, which would indicate no change in attentional bias. Error bars represent the 95% confidence intervals around mean change in attentional bias for each group at each timepoint.

Associations with Antidepressant Effect.

The association between antidepressant effects and change in AB from Study 1 and 2 are displayed in Figure 3. In Study 1, change in AB from pre- to post-infusion was not related to subsequent changes in MADRS scores (r=−.039, p=.895). However, in Study 2, change in AB was significantly related to subsequent changes in MADRS scores (r=.463, p=.030). This effect was maintained when treatment allocation (mindfulness training: yes, no) was included as a covariate (p=.033).

Figure 3. Correlations Between Percent Change in Attentional Bias and MADRS Scores following Ketamine Infusion in Study 1 and 2.

Due to age differences between participants in Study 1 and 2, post hoc linear regression analyses were conducted to examine age as a potential moderator in correlational analyses collapsed across samples. Age did not moderate the association between simultaneous changes in MADRS and AB scores, t(35)=−.248, p=.806.

Discussion

The present study provides evidence that intravenous ketamine rapidly and consistently ameliorates attentional bias for sad stimuli, suggesting that attentional processes may serve as a cognitive mechanism underlying the rapid-acting antidepressant effects of the drug. Findings from Study 1 revealed that our novel dual probe task demonstrated ‘good’ test-retest reliability across a one-month period, establishing its psychometric utility in probing attentional bias among depressed individuals. Moreover, the current study marks the first time that this novel task has been used in a sample of participants with depression. Our findings offer compelling evidence that the task consistently captures a form of attentional bias that is robustly associated with cross-sectional depressive symptoms across three independent time points. Our results also demonstrated a significant reduction in attentional bias in the 24 hours post-ketamine infusion in two independent studies (Study 1 and 2), indicating a rapid and replicable modulation of affective information processing. In Study 2, the observed changes in attentional bias were also correlated with improvements in clinician-rated depressive symptoms, independent of whether participants received additional mindfulness training, highlighting the clinical relevance of attentional bias as a potential target for understanding and augmenting the antidepressant effects of ketamine. In Study 1, the lack of a similar association between changes in attentional bias and clinician-rated depressive symptoms could be attributable to the relatively small sample size (n=15). Alternatively, compared to the 88% of participants taking psychotropic medications in Study 2, participants who received ketamine in Study 1 were not taking any concurrent psychotropic medications, representing the only noted dissimilarity in the constitution of the two samples other than age, which did not moderate findings across the two samples. Additionally, MADRS subfactors did not explain the discrepant findings (see Supplement). The observed association in Study 2 but not Study 1 may reflect the complexity of these relations, as well as particular vulnerability to Type II error in Study 1 due to the smaller sample size. In sum, because of the dearth of past research establishing ketamine’s influence on affective information processing, these findings contribute valuable insights into potential cognitive mechanisms involved in ketamine's therapeutic action and underscore the continued importance of examining attentional bias to inform novel treatment targets for depression.

The replicable reduction in attentional bias following a single infusion of ketamine sheds light on a potential intermediary mechanism contributing to its rapid antidepressant effects. Initial interest in examining attentional bias as a causal mechanism underlying treatment response followed observations that SSRIs reduced attention to negative stimuli, preceding reductions in mood(31). Given that attentional bias shapes a patient’s tendency to selectively attend to stimuli that evoke sadness, it is theorized to contribute to the maintenance and persistence of depressive mood states(13). Neuroplasticity induced by a single ketamine infusion may lead to a window of malleability in affective information processing(32, 33), allowing patients to break entrenched patterns of attentional processing and learn to instead direct their attention toward more benign stimuli. Consequently, the interplay between ketamine-induced neuroplasticity and alterations in attentional bias may represent a critical mechanistic pathway through which ketamine exerts its rapid antidepressant effects. Additionally, other factors such as mood elevation and subjective experiences such as awe(34) may also contribute. Further research will be needed to further dissect the temporal relations between ketamine delivery, attentional bias, and drivers of the amelioration of depressive symptoms. Future studies with additional affective information processing assessments are also needed to determine whether changes in attentional bias are a specific mechanism of ketamine’s rapid-acting antidepressant effect or reflect a broader cognitive-affective shift.

A key limitation of ketamine’s clinical impact lies in the rapid dissipation of its antidepressant effects following a single infusion(35). Thus, there has been considerable interest in identifying adjunctive therapies to extend response(7). For example, a recent study demonstrated that pairing automated self-association training (ASAT), a form of cognitive bias modification (CBM)(36), with a single dose of intravenous ketamine significantly extended ketamine’s rapid antidepressant effects(12). In the current study, the consistent reduction in attentional bias post-ketamine infusion, observed across two independent studies, suggests that modulation of attentional biases could also serve as a potential adjunctive CBM strategy to enhance ketamine’s durability. Traditional forms of attention bias modification training (ABMT) have relied on the dot probe task, which has been criticized for its low reliability and ecological validity (21, 37). Our novel dual probe task offers a more reliable and ecologically valid measure of attentional bias, as demonstrated by its 'good' test-retest reliability and associations with depressive symptoms across three independent time points. These improvements may enhance the efficacy of ABMT interventions when paired with ketamine, providing a more robust method to extend its antidepressant effects. This is notable given that prior studies found no changes in attentional bias using the dot probe task following ketamine administration(23, 24). Improved reliability and task design may explain why the novel dual probe task detected a treatment effect whereas the dot probe task did not(20-22), though direct comparisons are needed to rule out confounding factors like sample characteristics and size. In sum, the study provides initial evidence for the novel dual probe task as a reliable and valid measure of depression-relevant attentional bias and suggests that it could be pivotal in developing novel ABMT paradigms that could be delivered alongside intravenous ketamine to optimize treatment outcomes.

The current study exhibits notable strengths in its psychometric emphasis, addressing a critical gap in past research. Previous investigations into ketamine’s impact on neurocognitive mechanisms have often lacked replication in independent patient samples, and assessments of these neurocognitive changes have typically used measures with uncertain or suboptimal psychometric properties. In contrast, results from our study establish initial evidence of 1) test-retest reliability, 2) convergent validity, and 3) replicability. First, the novel dual probe task displayed ‘good’ test-retest reliability across three time points in a one-month period, a critical prerequisite for measuring intervention outcomes. The test-retest metric, relative to other reliability indices, holds particular significance in a patient sample, as it provides a measure of the stability of the entire paradigm over time and over repeated assessments. Second, attentional biases were consistently and robustly associated with clinician-rated depressive symptoms across three distinct test-retest intervals in Study 1 that occurred before ketamine infusion, providing preliminary evidence of the task’s validity in assessing depression-relevant attentional biases. Finally, we replicated post-ketamine effects on the reduction of attentional bias toward sad films in an independent sample, indicating a generalizable effect that will likely apply to future cohorts of treatment-seeking patients with moderate-to-severe depression. Of note, although attention biases were reliably correlated with depression severity in this study, variability in bias scores (i.e., some falling below 0.5 despite elevated symptoms) suggests that our novel task also captures the expected heterogeneity observed in depression. Together, these findings underscore the clinical utility of the novel dual probe task as a reliable and valid tool for measuring depression-relevant attention biases. This tool also holds promise for advancing our understanding of cognitive mechanisms in depression and may inform the development of novel ABMTs that could be paired with ketamine infusions in future treatment studies.

The study had several limitations. First, although the dual probe task utilized continuous video clips, offering more realistic stimuli than static images, these commercially sourced clips may not fully represent authentic real-world content and could have introduced familiarity biases that differed across conditions. Future studies might use paired benign and sad clips from the same films or develop novel stimuli. Second, the modest sample sizes in both studies may have limited detection of small effects. Third, the lack of a placebo comparison (e.g., saline infusion) and a non-depressed comparison group limited our ability to rule out expectancy bias or determine depression-specific effects. Fourth, we only examined attentional biases 24-hours post infusion, leaving questions about the durability of these improvements. Fifth, task attrition due to low accuracy or position bias may limit generalizability. Suggestions to improve task engagement are included in the Supplement. Additionally, other affective information processing biases (e.g., memory biases) were not measured, and future work is necessary to determine the specificity of ketamine’s effects on attention. Finally, participants were required to have moderate-to-severe depressive symptoms but not treatment-resistant depression, which may limit the generalizability of our findings to the population most likely to receive ketamine. Future research with larger, treatment-resistant samples, placebo and case control groups, and extended follow-ups is essential to address these limitations and clarify the relations between attentional bias and ketamine’s antidepressant effects.

In conclusion, our study provides robust support for the impact of a single infusion of ketamine in the amelioration of attentional bias to sad films among participants with moderate-to-severe depression. The observed reduction in post-ketamine attentional bias suggests that ketamine exerts a modulatory effect on affective information processing in depression, offering novel insights into its antidepressant effects. The current study also introduces a novel dual probe task that demonstrates ‘good’ test-retest reliability and convergent validity over a one-month period, which provides a psychometrically robust measure for probing depressogenic attentional biases in future studies. Together, these findings align with emerging research highlighting the potential synergy between ketamine and adjunctive CBM interventions such as ASAT(12) and ABMT(7). The combination of pharmacotherapies aimed at enhancing neuroplasticity with targeted cognitive interventions that capitalize on these short-term windows of malleability highlights a promising avenue to optimize treatment outcomes and extend ketamine’s rapid acting effects. This synergistic approach holds significant promise for novel treatments for depression and warrants further exploration in future research.