Although an akinetic rigid syndrome is the hallmark of idiopathic Parkinson's disease (PD), striatonigral degeneration (SND), and progressive nuclear palsy (PSP), cognitive and behavioural dysfunctions are often found in these parkinsonian syndromes. If the neuropsychological deficits directly reflect underlying brain neuronal lesions, we may expect that the cognitive and behavioural changes associated with these syndromes depend on the specific pattern of subcortical or cortical lesions that characterise each disease. The earliest and major cognitive deficit in all three diseases is impairment of executive functions as shown by performance on tests sensitive to frontal lobe function. Is it possible to distinguish, among the deficits contributing to this subcorticofrontal syndrome, patterns that are more specific to one or another of the three diseases? This might be useful for clinical diagnosis, especially in early stages of the diseases when the signs often overlap, accurate diagnosis being important for devising investigational and therapeutic strategies. In the paper by Soliveri et al 1 in this issue (pp 313–318) patients with PSP performed worse than patients with PD and those with SND on most tests during their initial evaluation, whereas patients with PD and those with SND performed similarly. These results are in agreement with other recently published studies.2 3 Patients with SND had, however, reduced phonemic fluency compared with patients with PD. This difference has to be confirmed by other studies, but might be clinically relevant. Dementia was found only in PSP, and corresponds to a severe subcorticofrontal syndrome caused by lesions in the the striatothalamocortical pathways resulting in deafferentation of premotor and prefrontal areas.

Because the natural history of these disorders also varies widely, Soliveri et al attempted to enlarge the differential diagnosis by a longitudinal assessment. Unfortunately, after 2 years, significant proportions of patients with SND or PSP were too disabled or had died, confirming that these two diseases progress more rapidly than PD. Among the patients who could still be tested, those with PSP declined significantly more than the two other groups on the Wisconsin card sorting test, which evaluates conceptualisation and set shifting. Patients with PSP or SND, however, had worsened significantly more than patients with PD on the visual search test assessing attention and visual scanning. There was no correlation between motor and cognitive decline. The proportion of demented patients increased only in PSP. This is not surprising given the more diffuse neuronal lesions in this disease.

In conclusion, clinical studies indicate that a parkinsonian syndrome is the presenting feature of several different diseases. Although most patients with parkinsonism have an idiopathic Parkinson's disease, others have multiple system atrophy, progressive supranuclear palsy, corticobasal degeneration or Lewy body dementia. Neuropsychological assessment in the early stages of the disease can contribute to the clinical diagnosis.4 The time course and pattern of progression of cognitive and behavioural decline can also improve the diagnosis, as shown by the study of Soliveri et al,1 particularly when postmortem confirmation can be anticipated.5