Meta analytic studies of fluency paradigms in neuropsychology
Tests of verbal fluency are presumed to measure executive dysfunction yet in practice are sensitive to brain damage per se, and the comparability of fluency based on phonemic and semantic criteria has been questioned. Relative to healthy controls, the effect size for each type of fluency was quantified for patients with focal cortical lesions, and using the random effects meta-analytic model, mean effects calculated for sub-groups stratified according to lesion location and laterality. Focal frontal patients were comparably impaired on both measures (rs = .48 and .47 respectively), but semantic fluency was relatively more dependent on temporal structures (.63), and associated with a deficit substantially in excess of the corresponding phonemic fluency deficit (.47). Thus, whilst both types of fluency place comparable demands on executive processes, semantic fluency is relatively more dependent on semantic memory. Phonemic fluency deficits qualified as differential deficits (i.e. they exceeded the averaged performance deficit across a range of other cognitive tasks) for frontal, but not non-frontal or specifically, temporal patients. Fluency deficits were then quantified for patients with Traumatic Brain Injury (TBI), dementia of the Alzheimer’s type (DAT), schizophrenia, Parkinson’s disease (PD) and depression. AS for frontal patients TBI was associated with comparable phonemic and semantic fluency deficits (rs = .46 and .43 respectively), and for moderate and severe injuries these deficits were differential deficits. For patients with DAT, schizophrenia and depression there was evidence of generalised cognitive dysfunction, but for each disorder semantic fluency was more impaired than phonemic fluency. However, differential deficits on tests of semantic memory only emerged when there were also substantial demands on effortful retrieval and/or cognitive speed. Finally, both demented and non-demented PD patients were also more impaired on semantic relative to phonemic fluency. However, semantic memory was impaired even when demands on effortful retrieval and cognitive speed were minimal.