A neural network model of peri-hand space representation and its plastic properties related to tool use

In the present study, we investigated the neural basis of peri-hand space representation and its plastic modifications related to tool-use by means of a neural network model. The model includes two networks (one per hemisphere/hand); each network includes two unimodal areas (visual and tactile, respectively) linked via synapses to a third downstream bimodal (visual-tactile) area. The two hemispheres interact via inhibitory synapses. The model assumes that synapses linking unimodal to bimodal neurons can be reinforced by a Hebbian rule during tool-use training; this reinforcement is also under the influence of spatial attentive mechanisms. To assess the effects of tool-use training, we simulated the visual-tactile interaction both in a healthy and a pathological subject with left tactile extinction, before and after the training. Results show that the integrative visual-tactile peri-hand area, limited around the hand before the training, is modified by tool-use to include new sectors of space, thanks to the sprouting of new visual synapses. The new size and shape of peri-hand area is determined by the attentive mechanisms active during tool-use training, which highlight only specific portions of the visual space functionally relevant to the use of the tool. The model may be of value to analyze the neural mechanisms responsible for representing and plastically shaping peripersonal space, and for the interpretation of psychophysical data on neurological patients with spatial perceptual deficits.