Numerical deficits in a single case of basal ganglia dysfunction

Abstract

The present investigation assesses specific numerical difficulties in a patient (SJ) with basal ganglia (BG) dysfunction. While previous studies on number processing in BG disorders typically tested arithmetic facts by production tasks, the present study uses production, recognition (verification, multiple-choice) and indirect (number-matching) arithmetic tasks. Patient SJ was severely impaired in production and to a lesser extent in verification and multiple-choice tasks. In number-matching, an abnormal latency pattern was found. This study extends previous research by indicating that BG dysfunction may not only affect production processes and sequencing, as was found in previous investigations, but may lead to a breakdown of semantic relationships of arithmetic facts.

Keywords:

• Arithmetic
• Verification
• Number-matching
• Fronto-subcortical
• Basal ganglia

Support for this research was provided by the Austrian FWF grant P18896-B05 and by the Medizinische Forschungsförderung Innsbruck PDO-Nr. 12880 (2007-419).

Notes

¹Following LeFevre et al. (1988), the arithmetic network's activation should be independent of the presence or absence of an arithmetic sign in the problem cue.

²Even when not specified, analyses were performed separately on arcsine-transformed error proportions and on ln-transformed correct RTs. Data were transformed to better approach the normal distribution. RT analysis was carried out after removing errors (wrong and missing answers) as well as outliers (i.e., RTs slower or faster than 2.5 SD from the mean). For each condition, missing RT values were replaced by the mean value of the condition.

³Unless specified otherwise, the comparison between the patient's and healthy participants' problem-size effect was performed in terms of slopes obtained by the regression of z-transformed correct RTs using target number as latency predictor.

⁴Within the control group, only one participant scored overall lower (62.5%) than the patient (66.7%) while rejecting false trials. This one participant made a considerable amount of errors when answering operand-related trials, whereas he performed as well as other control individuals on the remaining trial types. The exclusion of this control participant modifies the controls' range to 68.8–100%.

⁵In this case, there were differences among the controls' error variances; analyses were thus carried out on individual correlation coefficients.