Apathy symptoms modulate motivational decision making on the Iowa gambling task

Apathy as a syndrome manifests as reductions in motivation, goal directed thoughts, emotions, and behaviour [1]. The disorder is of frequent occurrence following neurological change [2]. Rehabilitation outcome studies on chronic patients going beyond 2 years post brain injury show that the majority of cases experience persisting apathy symptoms with related psychosocial problems characterised by lack of motivation, attenuated emotionality, decreased social contact and leisure activity, unemployment, marital problems and family breakdowns [3]. Apathy is also associated with significant caregiver distress and early institutionalization [4]. Despite these serious clinical implications, the nosological position of apathy and its associated neurocognitive profile remain poorly understood and appreciated in clinical practice [1].

Neurological models suggest apathy follows dysfunction in frontal-subcortical brain circuits crucial for motivation-related executive processing [5, 6]. The classic case of Phineas Gage provides the earliest documented case of such dysfunction. Despite his preserved capacities on basic cognition after his brain injury, Gage had significant socio-executive deficits including personality change and apparent apathy to his symptoms [7]. Similar cases have been reported elsewhere, e.g., [8‐11]. Apathy is often conceptualised as a dysexecutive syndrome [12], although studies have shown inconsistent results on the relationship between apathy and executive deficits assessed through standard executive function (EF) tests. Some studies have reported an association between apathy symptoms and poor performance on these tests [13, 14] while others have found no such an association. For instance, a review by van Reekum and associates [15] found a near middle split in the number of studies that reported a significant relationship between apathy levels and executive deficits (8 studies) and those that did not (7 studies).

Use of different tests and scales on different clinical samples across studies may partly account for the mixed results. Also, in some cases real-life dysexecutive behaviour has been seen to dissociate from deficits on standard EF tests [16‐18]. Patients with significant functional problems often perform relatively well on these ‘offline’ tests [19]. Robust tasks that are sensitive to real life socio-executive processing may help us understand the nature of neurocognitive impairment in apathy. The socio-executive processing needed for successful Iowa Gambling Task (IGT) performance is thought to analogue real life motivated behaviour [20, 21], and therefore makes a potentially useful test for the deficits in goal directed behaviour that comprise apathy symptoms.

A number of reasons make the IGT a plausible test for apathy symptoms. Firstly, brain areas thought to subserve emotional-executive processes crucial for effective IGT performance such as the medial frontal regions, anterior cingulate cortex, and amygdala form part of the cortico-subcortical neural circuitry responsible for motivated goal direct behaviour [22‐28]. Damage to these areas reliably causes apathy [29‐35]. Secondly, apathy and impaired IGT performance have both been linked to malfunctions of the dopamine system. Blockade of dopamine has been reported to impair, and stimulation of dopamine to improve, IGT performance [36]. Enhanced IGT performance due to stimulation of dopamine is also consistent with the view that dopamine mediates exploratory behaviour and the suggested effectiveness of dopamine agonists in combating apathy symptoms [37, 38]. With this converging evidence, the IGT provides an attractive premise to build a theoretical framework for understanding neurocognitive correlates of apathy symptoms.

In this study we assessed whether the IGT is sensitive to the presence of apathy symptoms in patients with acquired brain damage. It is also possible that the presence of apathy symptoms may impair performance on any test as a result of general motivational deficits. For that reason, we also obtained independent EF measures on the Brixton Spatial Anticipation test [39].

Tests for sphericity and homogeneity of variance were performed on the data to determine whether it met the assumptions for the use of parametric tests. Our data met these assumptions.

Patients with apathy made significantly more risky choices on the IGT compared to neurologically intact and brain damaged non-apathetic controls, both of whom (as groups) picked less from the risky decks as the game progressed. The shift from risky to the safe decks shown by healthy participants is in line with the performance patterns reported in similar studies [42‐47]. Non-apathetic brain damaged patients also shifted towards good decks as the game progressed.

It is possible that apathetic patients may perform poorly due to their general lack of motivation, and not necessarily due to a specific cognitive deficit related to the particular task. This argument is however inconsistent with our findings. For instance, putting in the Brixton test score as a covariate did not dilute the effect, suggesting the two patient groups’ performance on this test was not significantly different. We can claim with some degree of confidence that the IGT is sensitive to the presence of apathy symptoms, and distinguished apathetic from non-apathetic patients (and normal controls).

The IGT’s sensitivity to apathy symptoms makes it a potentially valuable instrument for both research and clinical practice. In a number of cases standard off-line EF measures fail to distinguish between patients with real-life socio-executive deficits like apathy and those without these deficits [48]. Standard EF tests may provide test takers with reliable task instructions and the type of feedback that is not found in more complex real life scenarios. On the other hand, the IGT has open ended behavioural choices and requires the generation of self-initiated choices under less explicit feedback. Such conditions provide a more valid analogue of socio-executive demands found in real life social environments. Our results suggest that apathy symptoms may arise from specific deficits in a patient’s capacity to produce or structure their own goal directed behaviour. This is in line with the realisation that apathetic patients rely heavily on others to structure their own activities [49]. Rehabilitation programmes in which daily tasks are structured, with enough prompts and cues may benefit patients with apathy. In this context, the IGT can be a useful addition to assessment batteries for detecting socio-executive deficits related to apathy.

Effective IGT performance is thought to depend on intact emotional processing [50]. According to this view (the somatic marker hypothesis), effective IGT performance benefits from subconscious emotional capacities that pre-bias responses away from bad behavioural choices. This bias enhances the selection and efficient execution of good behavioural choices, and is thought to underlie adaptive goal-directed activity, reasoning and decision-making in real life social contexts. It is tempting to suggest an inability to develop or fully utilise such emotional biasing signals to explain apathy symptoms. Empirical studies support this viewpoint. For instance, it has been observed that patients with apathy show deficits in matching emotional responses to social situations [51, 52]. Furthermore, there is evidence for a significant inverse association between apathy and autonomic excitation measured through heart rate reactivity [53]. Our study is however limited in the extent to which we can ascribe poor performance on the IGT to such lack of autonomic emotional inputs because we did not obtain concurrent measures of autonomic arousal.

The IGT’s sensitivity to apathy symptoms may also lie in its requirement for sensitivity to rewards and punishments. Part of the apathy symptom profile includes a dimension of indifference to rewards and punishments, lack of interest or concern, anhedonia, and social disengagement [54, 55]. Apathetic patients may have picked more risky cards due to insensitivity to punishments. Bechara and colleagues [56] have however demonstrated that sensitivity to rewards or insensitivity to punishment contingencies is not enough to explain impaired IGT performance. They either reversed rewards and punishments on the task such that good decks yielded higher immediate punishments but even greater delayed rewards, while risky decks gave off low immediate punishments and even lower future rewards, and in another condition increased the future disadvantages on risky decks. Patients with dysexecutive deficits preferred decks that had low immediate punishment to those with higher immediate punishment but more rewards in the long run. In other words large delayed rewards failed to lure the patients to the paying decks while on the other hand the patients were reluctant to choose from decks with huge initial punishment indicating that hyper sensitivity to rewards or insensitivity to punishments might not be a crucial factor on effective IGT performance.

Fellows and Farah [22] suggest that impaired goal directed behaviour can result from impaired future time perspective, which is a measure of an individual’s self defined future. Interestingly, deficits in future time perspective were found to correlate with self-reported apathy symptoms [22]. Whether this deficit can explain impaired IGT performance in this study is a subject for future research. There are also some suggestions that deficits in reversal learning or inhibition could account for impairments on the IGT. Effective performance requires that participants switch from initially high paying decks to lower paying but also low punishment decks. This involves reversal learning and the inhibition of responses to high reward/high cost decks - capacities that can be impaired in some brain damaged patients [57]. It is unlikely that deficits in reversal learning or inhibition could account for our results though, since apathetic patients were not distinctly impaired on the Brixton test.

Other factors, such as the effects of working memory deficits on IGT performance [58], and the cognitive impenetrability of the task [59, 60] have been suggested to explain impaired performance on the task. However, Bechara, et al. [61] have shown that working memory and decision making are dissociable and also shown that the majority of normal participants begin avoiding risky decks before confessing awareness of the game’s rewards and punishments contingencies. See also [62].

The IGT’s main weakness is its lack of specificity. A majority of different patient populations have shown deficits on the measure. Impaired IGT performance has been reported in schizophrenics [63‐65]; in pathological gamblers [66]; in alcohol, marijuana, and substance dependent individuals [67‐69]; in anorexics [70] in patients with intermittent explosive disorder and conduct disorder [71, 72]; in Huntington’s disease sufferers [24]; in traumatic brain injury patients [73]; in frontotemporal dementia [74] and in HIV infected patients [75]. It is nevertheless important to note that apathy is also a common syndrome in most of these disorders [15].

The IGT’s lack of specificity across different neuropsychiatric disorders could also be an indication of the task’s sensitivity to a wider spectrum of deficits, for instance, involving working memory, the patient’s interest and concern about the game, the processing of reward and punishment, explorative behaviour, and capacities related to the inhibition of choices on highly paying but risky decks.

The IGT’s sensitivity to apathy symptoms makes it a potentially valuable neurocognitive tool for assessing these symptoms. However the future utility of the task rests upon our ability to isolate the potential causes of impaired performance in specific clinical populations and syndromes. This should help unravel the individual aspects of the task that, for instance, give it sensitivity to apathy symptoms. Future studies may try and isolate or control for these various variables that can potentially impair IGT performance. Future studies may also consider using the electronic version of the IGT which allows the administration of the full 100 rounds even with poor or extremely good performers.