The contributions of lesion laterality and lesion volume to decision-making impairment following frontal lobe damage

https://doi.org/10.1016/S0028-3932(03)00081-2Get rights and content

Abstract

Lesions to prefrontal cortex (PFC) in humans can severely disrupt everyday decision-making, with concomitant effects on social and occupational functioning. Forty-six patients with unilateral lesions to prefrontal cortex and 21 healthy control subjects were administered three neuropsychological measures of decision-making: the Iowa Gambling Task, the Cambridge Gamble Task, and the Risk Task. Magnetic resonance imaging (MRI) scans were acquired from 40 patients, with region of interest (ROI) mapping of prefrontal subregions. The frontal patients showed only limited damage in medial and orbital prefrontal cortex, but greater damage in lateral prefrontal regions of interest. Patients with right frontal lesions preferred the risky decks on the Iowa Gambling Task, and differed significantly from left frontal and control subjects. Within the right frontal group, the preference for the risky decks was correlated with the total lesion volume and the volume of damage outside of the ventromedial prefrontal region. Right and left frontal groups did not differ significantly on the Cambridge Gamble Task or the Risk Task, and performance was not associated with lesion volume. The results indicate a laterality effect on the Iowa Gambling Task, and the contribution of prefrontal regions outside the ventromedial region to task performance. The Cambridge Gamble Task and Risk Task were less sensitive to the effects of unilateral frontal lobe lesions, and may be more selectively associated with ventral prefrontal damage.

Introduction

Decision-making operates in conditions of response uncertainty, and entails the evaluation of the reward and punishment contingencies to each of the available response options, followed by the calculation of the optimal option (Damasio, 1994, Rolls, 1999). Subtle but systematic errors in the ability of humans to reach the mathematically optimal decision have been well documented (e.g. the Gambler’s fallacy (Tversky & Kahneman, 1974) or the Wason selection task (Wason & Johnson-Laird, 1972)). However, qualitatively more severe decision-making deficits have been reported following brain injury, particularly where there is damage to the frontal lobes (Eslinger and Damasio, 1985, Satish et al., 1999, Shallice and Burgess, 1991). The nature of these deficits and their neuroanatomical substrates are the focus of the present study.

Bechara, Damasio, Damasio and Anderson (1994) developed the Iowa Gambling Task to assess decision-making in these patient populations. The task emphasises the contribution of emotional processing to decision-making. It requires the subject to learn the associations with reward and punishment of four card decks, in order to earn pretend money. Patients with bilateral lesions to ventromedial prefrontal cortex (PFC) were shown to be impaired on the task (Bechara et al., 1994). These patients, who also display poor decision-making in everyday life, persist in selecting from the ‘risky’ decks characterised by large immediate rewards but larger long-term punishments. Patients with damage to dorsolateral PFC were reported to perform within the range of healthy controls on the task (Bechara, Damasio, Tranel, & Anderson, 1998). Recent findings in a small number of patients with unilateral lesions have indicated that right-sided damage to ventromedial PFC is sufficient to impair performance, whereas left-sided lesions to this area do not impair decision-making (Tranel, Bechara, & Denburg, 2002).

A number of factors may contribute to deficient performance on the Iowa Gambling Task. A patient may have a genuine preference for high-risk options, or may be unable to compute the probabilities in order to discriminate risky from safe. They may be hypersensitive to reward, or insensitive to punishment, or may have difficulty learning the reward and punishment associations. More general problems with executive function may also disrupt performance in a learning context, including poor working memory and attentional inflexibility. Two novel decision-making tasks have been developed subsequently at the University of Cambridge by Rogers et al., 1999a, Rogers et al., 1999b, to characterise in more depth the role of PFC in decision-making. Both tasks utilise an independent trial structure to reduce learning demands, and ensure that all information needed to make each decision is displayed to the subject, to minimise working memory load.

In the Cambridge Gamble Task (Rogers et al., 1999a), subjects initially make a simple probabilistic judgment between two mutually exclusive outcomes, and then place a bet on their confidence in that decision. Four studies to date have used the Cambridge Gamble Task in groups of patients with different forms of frontal pathology. We have demonstrated increased betting in the presence of normal probabilistic judgment in three groups: (1) patients with large frontal lesions including the ventral PFC (Manes et al., 2002), (2) patients with aneurysms of the anterior communicating artery, the blood vessel that supplies ventral and medial PFC (Mavaddat, Kirkpatrick, Rogers, & Sahakian, 2000), and (3) frontal variant fronto-temporal dementia (Rahman, Sahakian, Hodges, Rogers, & Robbins, 1999), which is characterised by a disinhibition syndrome reminiscent of ventral PFC damage (Gregory & Hodges, 1996). One further study has shown decreased betting in the presence of impaired probabilistic judgment, in patients with frontal lesions including the ventral PFC (Rogers et al., 1999a). Thus altered betting on the Cambridge Gamble Task is consistently associated with damage to the ventral PFC.

The Risk Task was adapted from the Cambridge Gamble Task for use in functional imaging designs (Rogers et al., 1999b). The task again employs a probabilistic decision between two mutually exclusive outcomes, but on each trial there are fixed bets associated with either decision. The less likely outcome is always associated with the higher bet to create a situation of reward conflict. In two PET investigations using the task in healthy subjects, the contrast of the decision-making condition minus a visuo-motor control task detected significant activations in ventral PFC, and these activations were predominantly right lateralised (Rogers et al., 1999b, Rubinsztein et al., 2002). The functional imaging data therefore support the human lesion studies indicating ventral prefrontal recruitment during decision-making.

The selectivity of the ventral prefrontal contribution to decision-making was questioned by our recent investigation (Manes et al., 2002), in which we administered the three decision-making tasks described above to a group of patients with unilateral lesions within PFC. Neuroradiological assessment divided the group into patients with discrete orbitofrontal (including ventromedial), discrete dorsolateral, and discrete dorsomedial prefrontal lesions, and a fourth group with larger lesions affecting both dorsal and ventral aspects of PFC. The patients with orbitofrontal lesions performed in the control range on all three tasks, with the exception of deliberating for longer on two tasks. The group with large PFC lesions, in contrast, preferred the risky decks on the Iowa Gambling Task, placed higher bets on the Cambridge Gamble Task, and chose the less likely, but more rewarding, option more often on the Risk Task. However, lesion laterality was confounded in this study. In the orbitofrontal group, four of five patients had left-sided lesions, whereas in the large lesion group, four of five patients had right-sided lesions. It was therefore not possible to ascertain whether the large lesion group showed impaired decision-making because of the size of their lesions or the laterality of their lesions.

The aim of the present study was to extend the findings of Manes et al. (2002) by distinguishing the contributions of both lesion size and lesion laterality to decision-making. Patients with unilateral frontal cortex lesions were recruited from the Cambridge Cognitive Neuroscience Research Panel (CCNRP), and were combined with the Manes et al. (2002) group to provide a large group of 46 patients. The contribution of lesion laterality was examined using a three-way comparison of left frontals, right frontals, and healthy controls. The effect of lesion size was examined by correlating neuropsychological performance with lesion volume, measured by lesion tracing of magnetic resonance imaging (MRI) scans. A region of interest (ROI) analysis enabled the examination of decision-making performance in relation to the volume of damage in discrete prefrontal subregions (Aron, Fletcher, Bullmore, Sahakian, & Robbins, 2003). The markers of decision-making impairment were hypothesised to be: (1) increased selection from the risky decks on the Iowa Gambling Task, (2) increased betting on the Cambridge Gamble Task, and (3) increased choice of the less likely but higher rewarding option on the Risk Task.

Section snippets

Subjects

Patients were recruited from the Cambridge Cognitive Neuroscience Research Panel at the MRC Cognition and Brain Sciences Unit. The CCNRP is an accumulating database of volunteers with focal brain lesions. Forty-six patients with unilateral lesions in the frontal lobes were selected from the panel, for neuropsychological assessment. Twenty-four lesions were right-sided, 22 were left-sided. Lesion aetiology was tumour resection (24), haemorrhage (13), infarct (8), or abscess (1), and patients

Results

MRI lesion volumes were acquired from 19 left frontal and 21 right frontal patients. Mean volume of damage (in cm3) in the each of the regions of interest is presented in Table 2. Lesion volume was significantly larger in the right frontal group (mean 72.8 cm3, S.D. 60.0) than in the left frontal group (mean 35.1 cm3, S.D. 38.9) (t38=2.33, P=0.025). The right frontal group had significantly more damage in the middle frontal gyrus ROI only (t38=2.63, P=0.012). Therefore to further separate the

Discussion

The present investigation sought to dissociate the contribution of lesion size and lesion laterality to deficits in decision-making in patients with unilateral frontal lobe lesions. Three neuropsychological measures of decision-making were employed: the Iowa Gambling Task (Bechara et al., 1994), the Cambridge Gamble Task (Rogers et al., 1999a) and the Risk Task (Rogers et al., 1999b). Patients with right frontal lesions demonstrated severe impairment on the Iowa Gambling Task. They selected

Acknowledgements

Funding was provided by a Wellcome Trust programme grant to TWR, B.J. Everitt, A.C. Roberts, and BJS, and the research was conducted within the MRC Centre for Behavioural and Clinical Neuroscience. The authors would like to thank Dr. A. Bechara (University of Iowa) for use of the Iowa Gambling Task, Dr. P. Fletcher (University of Cambridge) for use of the ROI templates, and Adam Aron, Mike Aitken, Roshan Cools, and anonymous reviewers for helpful comments. We are also grateful to the members

References (25)

  • C.A. Gregory et al.

    Clinical features of frontal lobe dementia in comparison to Alzheimer’s disease

    Journal of Neural Transmission

    (1996)
  • Howell, D. C. (1997). Statistical methods for psychology. Belmont: Wadsworth Publishing...
  • Cited by (254)

    • Neural response to stress differs by sex in young adulthood

      2023, Psychiatry Research - Neuroimaging
    View all citing articles on Scopus
    View full text