Coalescing stochastic processes in retrival from semantic memory

Semantic memory retrieval is one of the most fundamental cognitive functions in humans. It is not fully understood and researchers from various fields of science struggle to find a model that would correlate well with experimental results and help understanding the complex background processes involved. To study such a phenomenon we need a relevant experimental protocol which can isolate the basic cognitive functions of interest from other perturbations. A variety of existing medical tests can provide such information, and the one we analyze  is the Category Fluency Test (CFT).  It was originally designed to measure frontal brain lobe  damages in  injured patients, and it tests directly the semantic memory retrieval, which is affected in cases of injury but can be also influenced  by dementia, Alzheimer syndrome, or just aging. This paper introduces a new paradigm in analysis of  the temporal structure of  CFT responses  by utilizing  coalescent  stochastic process model. We believe that this particular  model is relevant  to how  this cognitive function  operates and can lead to a better understanding of the background processes. The method turns out to be better  at separating the two cognitively different  groups studied here than the Weibull model from our previous paper Meyer et al.(2012), and could potentially be used for early diagnostics of dementia or Alzheimer's disease. Two other models, one based on the concept of Levy processes,  and one related to  the fractional Poisson model,  are also explored.