Tower of Hanoi and London performance after epilepsy surgery to the temporal or frontal lobes

The Tower of Hanoi (ToH) and Tower of London (ToL) puzzles are well established tools for analyses of planning and executive functions. There is, however, no consensus whether both puzzles measure identical cognitive skills and which brain areas are involved. This may be partly due to the various versions of ‘Towers' used in clinical and research environments. We used the tablet PC version of Hinz et al. (2007) that allows detailed analysis of e.g. different time measures, solution pathways, rule-breaks and errors and examined the performance of patients with well defined brain lesions after epilepsy surgery.

We tested 21 subjects with either left (n=8) or right temporal (n=6) and left (n=4) or right frontal (n=3) resection areas of at least 2cm diameter and compared their performance to that of 39 healthy controls. Each subject successively completed 20 ToH and 20 ToL tasks with difficulty graded according to minimum number of moves required (2 tasks each with 2 and 3 moves, 4 tasks each with 4, 5, 6, and 7 moves).

Controls as well as patients needed significantly more time to solve ToL than ToH problems (p<0.5). Patients required more total solution time for ToH than controls (p<0.5). They did not differ in their performance of ToL. The analysis of each single task (ToH & ToL) showed that patients in general did not need more time or moves to solve one task. Only three ToH tasks showed significant differences for initiation time, completion time, total solution time and number of moves for controls and patients: one four-move, one six-move and one seven-move problem (p<0.5). We found no systematic explanation why these three ToH problems were more difficult for patients. Lesion site in particular did not predict differences in performance. Still, total solution time by patients after left frontal resections differed by more than 2 standard deviations (SD) for ToH tasks and more than 1 SD for ToL from controls. Patients with other lesion sites showed variations of less than 1 SD.

Our data support Shallice's 1982 observation of a particular contribution of the left frontal lobe to ToL performance. In addition, ToH appears to be superior in the differentiation between patients and controls, since it covers a wider range of aptitude. In contrast, ToL seems to be more difficult and accordingly produces a ceiling effect.