Transcranial electric stimulation and neurocognitive training in clinically depressed patients: A pilot study of the effects on rumination

Highlights

• Depressive brooding was reduced following the Working Memory Training (CCT).

• Changes in working memory were related to changes in brooding after the WMT.

• tDCS did not influence this relation between working and brooding.

Abstract

Rumination is a cognitive–affective thinking style that plays a key role in the onset and maintenance of depression. Recently, it was shown that clinically depressed patients who received a neurocognitive training – involving two weeks of repetitive cognitive control exercises that necessitate prefrontal engagement – are more able to control over ruminative negative thoughts than patients who only received treatment as usual. Transcranial Direct Current Stimulation (tDCS) is a biological technique that can directly modulate prefrontal excitability via the manipulation of neural membrane potentials. In this randomized double-blind trial, we investigated whether bifrontal tDCS (anode over the left/cathode over the right dorsolateral prefrontal cortex (DLPFC)) would enhance the influence of a neurocognitive training on depressive brooding, the maladaptive form of rumination. Major depressed patients were trained using a procedure based on the Paced Auditory Serial Addition Task (PASAT), a task that relies heavily on working memory and is found to engage the DLPFC. One group (n = 19) completed the PASAT training together with active tDCS and another group (n = 14) completed the same training together with sham (placebo) tDCS. In both groups, depressive brooding was reduced following the PASAT training. Moreover, we observed that improvement in working memory over the course of the training was associated with a greater reduction in depressive brooding post- versus pre-intervention. However, tDCS did not moderate this association between changes in working memory and changes in depressive brooding. Possible explanations for this absent moderation of tDCS, as well as avenues for future research to influence ruminative thinking in depression, are discussed.

Introduction

Rumination, recurrent uncontrollable thoughts united by a common theme (Martin and Tesser, 1996) such as the possible causes, meanings and implications of negative mood states (Nolen-Hoeksema and Morrow, 1991), is a crucial cognitive–affective thought process in depression. This meta-cognitive thinking style makes individuals vulnerable to depression by maintaining and exacerbating depressive symptoms, and even by predicting the likelihood of recurrent depressive episodes (for a review, see Nolen-Hoeksema et al., 2008). Ruminative thoughts are associated with cognitive mechanisms such as impaired disengagement from negative representations and updating in working memory (e.g., De Lissnyder et al., 2012), and also to a neural dysregulation in frontocingulate and limbic circuits (for reviews, see Koster et al., 2011, Pizzagalli, 2011). More specifically, emotional stimuli activate the limbic circuit (Zald, 2003) which signals to the frontocingulate circuit to adjust the distribution of cognitive resources and, in turn, reduce the limbic activity (Hopfinger et al., 2000). However, this interplay between the neural activity related to emotional reactivity and cognitive control seems to be impaired in patients with major depression (Holmes and Pizzagalli, 2008), and appears to result in a maladaptive regulation of emotions (Davidson et al., 2002) and rumination (Koster et al., 2011). In sum, working memory processes (e.g. disengagement and updating of information) and the specific neurobiological functions associated with these processes have been proposed to be the mechanisms underlying the occurrence of ruminative thoughts and depression (for a review, see De Raedt and Koster, 2010).

In line with these process-oriented theoretical models in depression and rumination, non-pharmacological neurocognitive training procedures have been developed, during which depressed patients are repeatedly exposed to cognitive tasks linked to the engagement of prefrontal activity. Siegle et al., 2007, Siegle et al., 2014 developed a Cognitive Control Training (CCT) during which participants are trained on two neurocognitive tasks. One task of the CCT, a low-load version of sustained-attention training exercises are used in which patients are asked to focus their attention on external stimuli (e.g., bird sounds) (attention training: Papageorgiou and Wells, 2000, Wells, 2000). This latter task is meant to enhance selective attention to stay on the task when automatic ruminative thoughts could occur. In the other task of the CCT, the Paced Auditory Serial Addition Task (PASAT, Gronwall, 1977), working memory is trained and – as the demands on working memory are high – this is associated with more emotional reactions (e.g., frustration, negative thoughts, small amount of negative affect). As a result, working memory is trained in an emotional task context, which suggests that both the frontocingulate and limbic circuits are activated. Clinically depressed patients who received daily sessions (for two weeks) of this latter CCT showed a greater decrease in rumination than patients who had only received treatment as usual (Siegle et al., 2007, Siegle et al., 2014). Moreover, using functional magnetic resonance imaging (fMRI), patients demonstrated enhanced prefrontal activity during a digit sorting task and decreased amygdala activity during a personal relevance rating task after (as compared to before) the CCT (Siegle et al., 2007). So, the CCT in depressed patients influenced activation in neural correlates of the frontocingulate–limbic circuit, but also resulted in reduced rumination and depressive symptoms. Interestingly, as observed by Siegle et al. (2014), the CCT seemed to be most effective to reduce rumination when patients engaged in the working memory task by exerting cognitive resources at the start of the training (this was measured by pupillary responses to index task-related resource allocation). Furthermore, in a recent review, De Raedt et al. (submitted for publication) suggested that deficient prefrontal functioning in currently depressed patients might limit the effects of neurocognitive training. In other words, greater prefrontal engagement would augment the effects of neurocognitive training.

To directly enhance prefrontal excitability, transcranial Direct Current Stimulation (tDCS) can be used. This biological technique induces small changes (< 1 mV) in the membrane potential (Datta et al., 2009), acting in the frequency of spike timing and modifying net cortical excitability (Purpura and McMurtry, 1965), which can increase cortical perfusion and functional activity (Keeser et al., 2011). Anodal stimulation is found to increase cortical excitability, whereas cathodal stimulation is found to decrease excitability. Anodal tDCS of the prefrontal cortex causally enhances cognitive processes such as working memory (e.g., Fregni et al., 2005; for a review see Brunoni and Vanderhasselt, 2014) and conflict monitoring (Vanderhasselt et al., 2013a). Moreover, anodal tDCS of the prefrontal cortex has been found to reduce state rumination via a beneficial change in working memory processes (Vanderhasselt et al., 2013b) and also causally reduce other depressive symptoms (e.g., Brunoni et al., 2013a). Most important, tDCS doesn't require anesthesia and is well tolerated, which makes it a technique suitable to be combined with cognitive training (De Raedt et al., submitted for publication). It has also been demonstrated that concurrent neurocognitive training enhances the antidepressant outcome of anodal tDCS of the left dorsolateral prefrontal cortex (DLPFC) (Segrave et al., 2014). These findings strengthen the idea that the results of neuromodulation are better when anodal tDCS is delivered to a cortical region that is functionally active during a cognitive task. We recently reported that depressive symptoms are reduced after two weeks of training using the PASAT (see above): concomitant neurocognitive training and anodal tDCS of left DLPFC (cathodal over the right DLPFC) had beneficial effects in reducing depressive symptoms in older patients and those who perform better on the PASAT throughout the training (Brunoni et al., 2014b). However, the effects of two weeks of concomitant prefrontal neuromodulation and PASAT training on rumination – a core vulnerability process in depression – have not been reported so far.

Hence, the present study was designed to train clinically depressed patients repeatedly on working memory processes that engage the prefrontal cortex while anodal tDCS or sham (placebo) neuromodulation of the left DLPFC was administered. The aim of this study was to investigate the specific effects on rumination. In the studies of Siegle and colleagues, as was described above, the PASAT training was combined with other attention training exercises (Papageorgiou and Wells, 2000). However, using two tasks makes it impossible to disentangle the specific contribution of each task. Given that the PASAT is specifically known to activate the left middle frontal gyrus (including the DLPFC) (Lazeron et al., 2003), we only used this latter computer-based program to train working memory. Importantly, ruminative thoughts are associated with impaired processes in working memory (e.g., De Lissnyder et al., 2012). We assessed rumination with the Ruminative Responses Scale (RRS; Treynor et al., 2003), which consists of two subscales. The depressive brooding subscale assesses the degree to which individuals passively focus on depressive symptoms, the reasons for their distress, and a passive comparison of one's current situation with some unachieved standards. The reflective pondering subscale assesses neutrally valenced pondering and is considered to be a more adaptive form of rumination. Depressive brooding, the maladaptive self-critical component of rumination (Treynor et al., 2003) has been found to be specifically related to cognitive control impairments in working memory (De Lissnyder et al., 2010), and the activation in the DLPFC and the posterior parts of the dorsal anterior cingulate cortex during cognitive control operations (Vanderhasselt et al., 2011, 2013b).

Based in prior research (Siegle et al., 2007, Siegle et al., 2014), our hypotheses were that:

• working memory performance on the PASAT will be improved over the course and following of the training, with larger improvements in the tDCS condition (as compared to sham);

• depressive brooding reports will be reduced following the PASAT training, with larger effects in the tDCS condition (as compared to sham);

• the improvement in working memory will be related to the reduction in depressive brooding scores pre- versus post-training;

• the association between changes in working memory and changes in depressive brooding will be stronger in the tDCS condition as compared to the sham condition.