Assessing neurocognitive function in psychiatric disorders: a roadmap for enhancing consensus

Abstract

It has been challenging to identify core neurocognitive deficits that are consistent across multiple studies in patients with Obsessive Compulsive Disorder (OCD). In turn, this leads to difficulty in translating findings from human studies into animal models to dissect pathophysiology. In this article, we use primary data from a working memory task in OCD patients to illustrate this issue. Working memory deficiencies have been proposed as an explanatory model for the evolution of checking compulsions in a subset of OCD patients. However, findings have been mixed due to variability in task design, examination of spatial vs. verbal working memory, and heterogeneity in patient populations. Two major questions therefore remain: first, do OCD patients have disturbances in working memory? Second, if there are working memory deficits in OCD, do they cause checking compulsions? In order to investigate these questions, we tested 19 unmedicated OCD patients and 23 matched healthy controls using a verbal working memory task that has increased difficulty/task-load compared to classic digit-span tasks. OCD patients did not significantly differ in their performance on this task compared to healthy controls, regardless of the outcome measure used (i.e. reaction time or accuracy). Exploratory analyses suggest that a subset of patients with predominant doubt/checking symptoms may have decreased memory confidence despite normal performance on trials with the highest working memory load. These results suggest that other etiologic factors for checking compulsions should be considered. In addition, they serve as a touchstone for discussion, and therefore help us to generate a roadmap for increasing consensus in the assessment of neurocognitive function in psychiatric disorders.

Keywords: Checking; Ignore-Suppress; Neurocognitive; Obsessive Compulsive Disorder (OCD); Serotonin reuptake inhibitors (SRIs); Working memory.

Figure 1

Ignore-Suppress Task Methodology. In Ignore (A), an instruction to Attend to either the red or blue words appeared followed 1 s later by a 6-word memory-set for 4.5 s, followed 3 s later by a probe word. In Suppress (B), the 6-word memory-set appeared for 4.5 s, followed 1 s later by an instruction to Remember only the words of one color (red or blue). This was followed 3 s later by a probe word. In both Ignore and Suppress, a “Yes” response to the probe was required if the probe word matched one of the words that was supposed to be in working memory, and a “No” response was required otherwise.

Figure 2

Ignore-Suppress Reaction Times. Box and whisker plots showing reaction time data (for correct trials only) on the three different trial types (Control, Lure, Valid) for Healthy Controls (HC) and OCD subjects (OCD), separately for the Ignore and Suppress tasks. Participants were significantly slower in the Suppress condition as compared to the Ignore condition (936.03 vs. 745.45 ms, respectively) (p < .001), and were significantly slower in the Lure trials (917.61 ms) than in both the Control (797.87 ms) and Valid (806.74 ms) trials (p < .001). The interaction between Condition and Trial Type was significant (in Ignore: Control = 732.36; Lure = 765.59; Valid = 738.39; in Suppress: Control = 863.39; Lure = 1069.62; Valid = 875.09) (p < .001). However, the main effect of Group was not significant, nor were any of the other interactions (Group × Condition; Group × Trial Type; Condition × Trial Type × Group) (all Fs < 1.1).

Figure 3

Ignore-Suppress Accuracy. Box and whisker plots showing accuracy (percent correct) on the three different trial types (Control, Lure, Valid) for Healthy Controls (HC) and OCD participants (OCD), separately for the Ignore and Suppress conditions. Participants were significantly more accurate in Ignore than in Suppress (96.02% vs. 89.32%, respectively) (p < .001). However, the main effect of Group was not significant (HC: 93.17%; OCD: 92.16%), nor was the main effect of Trial Type (Control: 94.05%; Lure: 91.27%; Valid: 92.68%), nor were any of the interactions (Group × Condition; Group × Trial Type; Condition × Trial Type; Condition × Trial Type × Group).

Figure 4

Exploratory Clinical Correlations. (A) Scatter-plot demonstrates positive correlation of increased doubt/checking with RT on Valid Suppress trials [Spearman rank correlation: p < 0.07; ρ= .43; r² = .248]. (B) Scatter-plot demonstrates negative correlation of increased doubt/checking with accuracy on Valid Suppress trials [Spearman rank correlation: p < 0.04; ρ= −.50; r² = .039]. (C) Scatter-plots demonstrate no significant correlations between hoarding, contamination/cleaning, taboo thoughts, or symmetry/ordering and RT on Valid Suppress trials (all p > 0.3).