The objective of the present work was to assess supra-second ITT performance in individuals with BD. Specifically, we were interested in whether BD disorder sub-type, depressive mood, or antipsychotic medication-use altered supra-second interval timing in our cohort of patients. Our results suggest that, although ITT performance and frontal theta were abnormal in the BD group compared to the CT group, within BD sub-groups there were no differences in performance or theta power. Together with previous work indicating that individuals with BD show impairments in supra-second (Mahlberg et al. 2008; Bschor et al. 2004), sub-second (Bolbecker et al. 2014a, 2009a), and implicit motor timing (Bolbecker et al. 2011), our work suggest that an altered ability to assess the passage of time is a fundamental cognitive abnormality in BD.

Cognitive abnormalities are pervasive in psychiatric disorders, although less studied compared to mood dysregulation. For BD alone, the abnormalities in timing described in the present work, are part of a larger set of cognitive difficulties also including irregularities in attention/working memory, executive function, verbal/visual memory (Andersson et al. 2008). Additionally, deficits in interval timing have been previously reported for other neuropsychiatric conditions including SCZ and Parkinson’s disease (Singh et al. 2019; Parker et al. 2017, 2013), indicating the importance of studying the cognitive abnormalities associated with such disorders.

Although cognitive impairments are widely reported by BD patients (King et al. 2019), few reports have attempted to triangulate which cognitive impairments may be state vs. trait characteristics of BD. Because clinical characteristics vary between BDI and BDII, and because previous work suggests a link between depressed mood/antipsychotic medication-use and altered temporal processing, we assessed if these characteristics varied within our group of BD patients.

First, our results indicate that supra-second timing performance is not altered as a function of BD disorder sub-type. The extent to which cognitive profiles differ between BDI and BDII is debated in the literature. While some work suggests that BDI presents with more significant cognitive impairments (Sole et al. 2012), other studies suggest similar cognitive profiles between the two sub-groups (King et al. 2019). Indeed, recent work suggests that BDII patients show impairments in attention/working memory, executive function, verbal and visual memory, and motor speed compared to neuronormative controls, just as BDI patients do (Andersson et al. 2008). Our results add to this body of literature, indicating that cognitive impairments in the supra-second interval timing domain, do not differ by BD disorder sub-type.

Additionally, our work suggests supra-second interval timing abilities do not differ between depressed and euthymic BD patients. The lack of distinction between these two groups is interesting given the well-established link between depression and a slowing of time perception in the supra-second domain (Kent et al. 2019). However, previous work suggests that 40–60% of euthymic BD patients may present with neurocognitive impairment (Sole et al. 2012). Indeed, work by Martino and colleagues (Martino et al. 2008) assessed six cognitive domains (attention, verbal memory, language, psychomotor speed, executive function, and facial emotional recognition) in BD and found that 62% of euthymic BD patients showed cognitive impairments, with 40% of patients showing 1 or 2 impaired domains, and 22% of patients showing impairments in 3 to 5 domains. Our findings thus suggest an additional domain—supra-second interval timing—where BD patients show impairments even in the euthymic state, adding to the growing literature indicating cognitive markers as a fundamental characteristic of the disorder.

Finally, our work suggests that antipsychotic medication-use does not alter supra-second timing in BD patients. Past work suggests a negative association between antipsychotic medication-use and IQ in BD (Abrams et al. 1981). Specifically, tests of working memory, set-shifting, and response initiation/inhibition are negatively affected by antipsychotic medication-use (Frangou 2005). However, not all cognitive measures in BD are affected by antipsychotic medication-use, including measures of response planning and general working memory. In the context of this literature, our negative findings concerning antipsychotic use and interval timing are surprising. However, the dependence of timing abilities on working memory and response planning, two cognitive features not altered by antipsychotic use, could explain these findings. Another possibility is that participants were on low antipsychotic doses. Because antipsychotic dose is related to the degree of cognitive impairment, this could explain the lack of group difference. However, dose information was not collected, thus this analysis could not be conducted leaving space for future work.

The present work identified abnormalities in frontal (Fz) theta oscillations during the ITT in BD patients compared to CT participants. Previous work suggests that ITT performance increases frontal theta power compared to rest. This pattern of activity was indeed detected for the CT group, where visual inspection of ITT vs. resting state graphs suggests that frontal theta power was higher during the ITT (Figs. 2E vs. 3A). However, this pattern was not detected for the BD group, where average power stayed approximately the same during task and rest. Frontal theta oscillatory activity is widely related to cognitive processing for a variety of tasks (Kahana et al. 2001; Amarante et al. 2017). Indeed, compelling theories even propose frontal theta oscillations as a mechanism by which cognitive control may be biophysically realized (Cavanagh and Frank 2014). In this light, our results can be interpreted to suggest a failure in the mechanisms subserving general cognitive functioning and attentional direction in the BD group, which then results in timing deficits.

Previous work suggests that, compared to neuronormative controls, patients with SCZ show abnormal frontal low frequency (delta + theta) activity during the ITT (Singh et al. 2019). Our work suggests that the relationship between abnormal frontal theta and impaired ITT performance may not be a characteristic of SCZ specifically, extending to BD as well. Finally, in SCZ patients, work suggest that abnormalities in theta power in the 500 ms window following timing-cue presentation is related to abnormal supra-second ITT performance (Singh et al. 2019). Although the primary objective of the present work was to analyze ITT performance and theta power within BD sub-groups, not between BD and CT groups, because of this previous SCZ work, secondary analyses were added to identify specific epochs of altered theta power during the task. These analyses were time-locked to cue presentation and response. Results suggest that oscillatory abnormalities in BD were not time-locked to the post-cue interval as they were in SCZ (Additional file 1: Figs. S2B-C and S3B-C). One surprising finding from the present dataset is that individuals with BD showed lower theta power surrounding the response for the long interval (Additional file 1: Fig. S3F), but not the short interval (Additional file 1: Fig. S2F). This parallels performance data where BD patients show altered precision estimates for the long interval (Fig. 2D [right]), but not the short interval (Fig. 2C [right]). These results could indicate that frontal theta power is more closely linked with precision than response accuracy. However, further work is necessary to substantiate this claim.

Using timing task performance to triangulate single regions which may be abnormal in BD presents a challenge, as the neuroanatomy of time processing is famously diffuse (Buhusi and Meck 2005) involving the coordinated functioning of multiple brain regions and neurotransmitter systems. One mechanism underlying the altered ITT performance observed in the present work may be the abnormal functioning of the dopamine system in individuals with BD. Indeed, the dopamine hypothesis of BD, which proposes intrinsic dysregulation of dopamine receptor transporter homeostasis (Ashok et al. 2017; Wittenborn 1974), is widely used to explain the pathophysiology of this disorder. Additionally, in other disorders where dopaminergic pathway function is altered, such as SCZ, Parkinson’s, or Huntington’s disease, abnormalities in temporal processing have also been reported (Buhusi and Meck 2005). However, the absence of an effect of antipsychotic treatment on ITT performance weighs against the interpretation of timing deficits being caused by dopaminergic system abnormalities, as this medication class primarily targets the dopamine system.

Another possible mechanism subserving the ITT performance and frontal theta deficits identified in the present work is the well-characterized frontal cortical abnormalities observed in individuals with BD including reductions in cortical grey matter (Moorhead et al. 2007; Michel and Koenig 2018). Indeed, compromised frontal cortical activity has been linked to abnormalities in supra-second interval timing (Buhusi and Meck 2005). This suggests a suggests a failure in the frontal mechanisms subserving time perception in BD patients, expressed electrophysiologically as unaltered frontal theta power and behaviorally as impaired supra-second interval timing.

Although the present work suggests task-wide abnormalities in frontal theta power in the BD group compared to the CT group, it is not ideally set-up to answer when exactly these abnormalities occur. It is possible that abnormal frontal oscillations extend throughout the timing interval, or are limited to the cue, response, or feedback-integration intervals. Our analyses suggest abnormalities locked to the response epoch of the long interval, however, these analyses are preliminary and underpowered (Additional file 1: Figs. S3 and S4). In addition to abnormalities in power, it is possible that a weakening of phase coupling associated with the integration of timing cue processing translates to greater variability in response time. Future work should disentangle this issue.

The present sample is skewed towards BD patients, as the CT group comprises 20% (n = 6 or 7) of the total study population while the BD group comprises 80% (n = 24) of the population. Because of this, the precision of estimates where comparisons between CT and BD groups are made are limited. Indeed, our results should be carefully interpreted as a first step towards understanding supra-second ITT performance in BD, and further, larger studies are needed. However, two indicators within the present data-set point to the reliability of these findings (1) the variability within the control group for most analyses is quite small, smaller than the BD group, in all cases. And (2) we are not the first group to report differences in ITT performance between BD and CT groups, indeed our findings are confirmatory of previous work (Mahlberg et al. 2008; Bschor et al. 2004).

Another factor to consider within the present dataset is that polypharmacy was the treatment plan norm within the patient population, which complicates interpretation of how timing results vary by antipsychotic use. To disentangle this, we calculated an ACB score for each patient. Our regression analysis suggests a weak relationship between the SIT peak time index and ACB score only. However, ACB scores within our data were not normally distributed, and clustered between 1 and 5, complicating interpretation.

Finally, we were unable to assess how mania alters supra-second interval timing performance, as none of our BD patients were in a manic state. This question is of particular interest as results are not consistent within the literature: while some studies suggest that manic patients under-estimate supra-second intervals (Mahlberg et al. 2008), other suggests that manic patients over-estimate such intervals (Bschor et al. 2004). However, this remains an open question for future work.