Planning as a prototypical executive function is frequently compromised in clinical samples. Analyses of rule breaking during performance on tower tasks are highly informative for clinical inferences on planning deficits, but are as yet only available for manual task versions. Therefore, the present study investigated whether valid rule-break information can also be gained using a computerized Tower of London (TOL) version.
For patients with stroke (N = 60), Parkinson's disease (PD; N = 51), and Mild Cognitive Impairment (MCI; N = 29), rule breaks (e.g., selecting a blocked ball) and online-planning errors (re-considering the next move) on the TOL-Freiburg (TOL-F) were compared to that of matched healthy controls, and diagnostic accuracy for these measures was assessed.
Stroke and PD patients committed significantly more rule breaks and online-planning errors than matched controls, with this group difference being significantly greater for rule breaks. For MCI patients, only a trend for elevated numbers of rule breaks emerged. In all patients, rule breaks and errors increased as a function of problem complexity. Diagnostic accuracy was the highest for PD and lowest for MCI patients, with both rule breaks and errors demonstrating moderate sensitivity but higher specificity across all samples.
Direct attempts to break the TOL-F's task rules are particularly informative of clinical impairments in planning ability. Similar to findings from studies employing manual versions, using a computerized tower task revealed consistent rule break and error patterns across patients. Thus, computerized tower tasks can yield useful clinical information on rule-breaking behavior during planning in clinical populations.