This proposition does not assume segregation of cortical and subcortical systems but instead a striatal-cortical gradient for the implementation of control over stimulus and additional abstract response set representations. One of the strengths of the present study is that the patient groups were matched in terms of age and IQ to a single control group. This is important for any formulations based on comparative

neuropsychology, due to the cognitive and neural effects of ageing. The frontal lobe hypothesis of ageing holds that older adults display disproportionate deficits on tasks of cognitive control reliant on frontal function (Dempster, 1992; West, 1996). Age-related reductions in grey matter volume EPZ-6438 nmr (Coffey et al., 1992; Raz et al., 1997) and metabolism (Azari et al., 1992; Salmon et al., 1991) are greater in prefrontal than sensory cortical regions and manifest as significant deterioration in performance on tasks supported by frontal regions, compared with relatively smaller deficits on non-frontal tasks (Ardila & Rosselli, 1989; Daigneault, Braun, & Whitaker, 1992; Shimamura & Jurica, 1994). In particular, relative to their younger counterparts, older adults demonstrate increased AZD0530 order SC (Cepeda, Kramer, & Gonzalez de Sather, 2001; Kramer, Hahn,

& Gopher, 1999; Mayr, 2001; Verhaegen, Kliegl, & Mayr, 1997), especially in the absence of external cues (Kramer et al., 1999; Kray & Lindenberger, 2000). Furthermore,

ageing-related switching deficits are associated with stimulus and response bivalency leading to increased task interference during task set reconfiguration (Mayr, 2001), that represents another set of conditions which engage the PFC. These specific effects of age can be assumed to affect the present groups equally, thereby lending further credence to the parallel examination of frontal and parkinsonian switching performance. The current design, analyses and findings focus on switch check details costs, an index of transition within each rule condition. However, an alternative interpretation of these switching dissociations as a function of whether a switch of task requires a reconfiguration in response rule could invoke differences in difficulty to account for the neural sensitivity of abstract rule switching: it was in fact predicted that applying these rules, at which both frontal lesion and stage II but not stage I PD patients were impaired, would produce longer RTs compared with concrete naming rules. Switching between the former rules could therefore be more sensitive to cognitive dysfunction. This account cannot be ruled out by the current data.