Authors:

WOOLGAR, A., BOR, D., DUNCAN, J., Parr, A., CUSACK, R., NIMMO-SMITH, I., Antoun, N, Manes, F. & Torralava, T

Reference:

Year of publication:

2007

CBU number:

6503

Abstract:

A large study of 82 patients with focal brain damage in diverse regions of the brain is presented. We relate lesion volume and location to a measure of deficit in fluid intelligence (Spearman’s g) calculated by comparing an estimation of premorbid and measured IQ. We go on to explore, with fMRI, the functionality of non-lesioned areas of the patient brain when performing non-verbal reasoning tasks.Previous WorkIn frontal and parietal cortex, a network of “multiple demand” (MD) regions is recruited in tasks of many different kinds (Duncan & Owen 2000). Recent imaging work in our unit has shown a striking similarity between this network and the pattern of activation produced when solving standard tests of g (Bishop et al, in prep). In the current study we investigate whether damage to these regions is predictive of a deficit in g.Fig.1 (a) MD regions (Duncan 2006) (b) activation to tests of g (Bishop et al in prep)Behavioural MeasuresTo assess the contribution of brain damage to a patient’s fluid intelligence score, we estimated the discrepancy between premorbid and measured g. Measured g for each patient was based on scores on two tasks: Cattell Culture Fair, a test of non-verbal reasoning and a standard measure of g; and Letter Sets, a further test of reasoning previously shown to be highly correlated with g. An equation for predicting scores on each of these tests from a participant’s age and performance on the National Adult Reading Test (NART) was derived from the scores of 25 age-matched healthy volunteers. Since scores on NART are highly correlated with tests of g in healthy volunteers and are also thought to be resistant to deterioration after brain damage, it was possible to predict premorbid scores on Cattell and Letter Sets for each of the patients. The discrepancies between the premorbid and measured scores on each of the tests were then combined to give a single measure of deficit.ResultsLesion volume was found to be the greatest single predictor of behavioural deficit. The correlation was most marked in patients with right frontal lesions and less marked for lesions elsewhere. The extent to which lesions compromised MD regions of the brain was also a good predictor of behavioural deficit, but the close correlation between this and lesion volume made the two factors hard to disentangle.Functionality of MD regions after brain damageThe extent to which better performance on these tasks was related to the functionality of MD regions of the damaged brain was explored using fMRI. In particular we were interested to explore why some patients showed large behavioural deficits while other patients with similar lesions showed little or none.We used an adapted version of the Cattell and Letter Sets tasks previously shown to produce reliable activation of MD regions. We subtracted activation recorded during blocks of easy (low g-loaded) problems from blocks of difficult (high g-loaded) problems. Comparison of activity across patients and controls shows how the damaged brain reacts to produce patterns of preserved or impaired behaviour.