A neurocomputational model of visuotactile interactions in the temporal domain in the case of unimanual and bimanual stimulations

The communication between the different senses is fundamental to be able to interact with our surroundings. In our everyday life, our senses capture and exchange information in a process called “multisensory integration”: in this way, we are capable of generating precise responses to various situations, overcoming the uncertainty and the ambiguity that could be present if we were to rely on a single sense. Multisensory integration occurs when stimuli of different modalities are presented together. If the stimuli are presented sequentially, a latency in the response occurs, due to the “cost” of switching between modalities: this phenomenon is called “Modality Switch Effect” (MSE). The aim of this work is to propose a neurocomputational model which, with high simplicity, tries to mimic the processes underlying the interaction between vision and touch, specifically the multisensory integration and the MSE. The starting points were two models previously proposed, which simulated the described effects for audio-visual and visuo-tactile pairings, as well as the effects of bimanual stimulations. The model’s basal behaviour was tested, then its performance was compared with experimental data, introducing noise at the inputs, as well as a variable ISI to replicate the unpredictability of the stimulation; the results showed that the network was successful in replicating the experimental results, even with some limitations due to the reduced amount of available data. Further research should be performed to collect more data and validate the model; moreover, a similar architecture could be exploited to further investigate the processes underlying the interaction between hemispheres.