Humans can quickly engage a neural network to transform complex visual stimuli into a motor response. Activity from a key region within this network, the intraparietal sulcus (IPS), has been associated with evidence accumulation and motor planning, thus implicating it in sensorimotor transformations. If such transformations occur within a brain region, a key and untested prediction is that neural activity reflecting both the parametric amount of evidence available and the timing of motor planning can be independently manipulated. To investigate these ideas, we constructed a dot motion discrimination task in which information about response modality (what to use) and response mapping (how to use it) was provided independently either before or after presentation of a dot motion coherence stimulus whose strength varied across trials. Consistent with our hypothesis, activity within IPS covaried with dot motion coherence during the stimulus phase, and as information necessary for the response was delayed, the peak of IPS activity shifted to the response phase. In contrast, areas such as the motion-sensitive region MT+ and the supplementary motor area demonstrated activity limited to the stimulus and response phases of the task, respectively. These results show that activity in IPS correlates with temporally dissociable representations consistent with both evidence accumulation and motor planning, and suggest that IPS is a core component for sensorimotor transformations within the perceptual decision-making network.© 2013 Elsevier Inc. All rights reserved.
IntroductionThe ability to link sensation with action is integral to even the most fundamental of decisions. In humans, a number of studies have identified a frontoparietal network involved in processes during perceptual decisions, including sensory processing, attentional control, evidence accumulation, and motor planning (Heekeren et al., 2008; Kayser et al., 2010a,b;Liu and Pleskac, 2011;Ploran et al., 2007Ploran et al., , 2011Rowe et al., 2010;Ho et al., 2009;Tosoni et al., 2008). Previously we demonstrated that for subjects performing a dot motion discrimination task, this network displays a blood oxygen level-dependent (BOLD) response that varies inversely with motion coherence (Kayser et al., 2010a). Consistent with previous and subsequent findings (e.g., Hebart et al., 2012;Ho et al., 2009;Kayser et al., 2010b), this negative parametric effect matched predictions of a proportional-rate diffusion model for evidence accumulation to a threshold (Palmer et al., 2005;Ratcliff and McKoon, 2008), and linked human brain activity with a computational model of the transformation from stimulus to response.Although this parametric effect is necessary, it is not sufficient to define regions that support such sensorimotor transformations. Human studies, for example, show that areas such as the motion-sensitive region MT+, supplementary motor area (SMA), and intraparietal sulcus (IPS) all display this parametric effect. Importantly, results from macaque studies show that ...