Abstract
The human neural correlates of perceptual decision-making bear a remarkable correspondence to the predictions of sequential sampling models. However, to date, the research focus has been on binary choices, while real-world decisions typically offer more than two possible responses. Here, we describe a transcranial magnetic stimulation (TMS) paradigm capable of tracking decision-related evidence accumulation in four effectors while human participants perform two and four-choice perceptual decisions. Thirteen participants collectively contributed around 54,000 trials on a random-dot motion discrimination task. After pre-processing, around 30,000 non-TMS trials were used to estimate the best-fitting parameters of sequential sampling models, and around 14,000 TMS trials were used to estimate smoothed motor-evoked potential (MEP) signals associated with different possible responses. Direct comparison of these MEP-based signals, which appeared to provide a downstream readout of decision-making computations, to the accumulation profiles predicted by the leaky competing accumulator (LCA) model revealed marked qualitative similarities between model predictions and corticospinal signals. We thus demonstrate that TMS can be used to track the multi-alternative decision variable, and show that, in line with model predictions and findings from non-human primates, the number of choice alternatives affects the baseline level of neural accumulation in humans.
Meeting abstract presented at VSS 2018