A PDP Simulation of the E ffects of Transcranial Magnetic Stimulation on Semantic Cognition

Transcranial magnetic stimulation (TMS) is a useful tool for determining cortical interactions that take place during semantic cognition (Pobric et al, 2007). TMS of the anterior temporal lobe (ATL) induces specificity-graded conceptual breakdown of stored representations resulting in differential effects on naming tasks depending on the level of specificity required. This paper aims to elucidate this effect using the Rogers et al (2004) model of semantic cognition. TMS effects are modelled by reducing the gain of the affected units to simulate reduction in neuronal sensitivity. Results show that basic-level naming is more robust to rTMS than subordinate level naming as found by Pobric et al. In addition the model predicts that superordinate naming should be even more robust to rTMS than basic level naming. This specificity-graded breakdown of semantic memory appears to be independent of learnt word frequency. This supports evidence that the ATL’s function is that of category differentiation. K eywords: neuroscience; psychology; cognitive architectures; concepts and categories; distributed cognition; memory; semantics; computational neuroscience; computer simulation; neural networks