As mentioned above, the entorhinal cortex has been implicated as an important structure in the ability to form associations between arbitrarily related items, i.e., items that do not have a pre-established semantic link between them. Early DAT patients' difficulty in forming such associations was demonstrated using the computerized Paired-Associate Learning (PAL) subset of the Visual Memory Battery of the Cambridge Neuropsychological Test Automated Battery (R. Morris, Evenden, Sahakian, & Robbins, 1987). Measuring visuospatial new learning, in the PAL task subjects are required to form an association between a gradually increasing number of objects and their hidden location on a computer screen (Fowler et al., 2002).
Fowler et al. (2002) in a longitudinal study found that the PAL task identified best subjects with a diagnosis of questionable dementia (QD) at base line who all met criteria for AD in a two year follow-up. In addition, the deteriorating subgroup was not distinguishable from a non-deteriorating subgroup of QD on more standard neuropsychological examinations. The authors concluded that the PAL task should be utilized as an independent clinical tool (Fowler et al., 2002).
Recently, Blackwell et al. (2004) in a similar study developed a predictive algorithm that identified perfectly a group of AD patients at 32 months follow up based on their performance on both the PAL task and a semantic task at base line (Blackwell et al., 2004). In summary, the PAL task appears to be a very sensitive screening tool for the earliest cognitive changes associated with pre-clinical AD.