Predicted sensory consequences of voluntary actions modulate amplitude of preceding readiness potentials

Highlights

• Sensory outcomes are represented in brain activity preceding voluntary actions.

• Amplitude of readiness potential (RP) is greater for actions with auditory outcomes.

• Greater RP amplitude is not related to differences in sensory expectation.

Abstract

Self-generated, voluntary actions, are preceded by a slow negativity in the scalp electroencephalography (EEG) signal recorded from frontal regions (termed ‘readiness potential’; RP). This signal, and its lateralized subcomponent (LRP), is mainly regarded as preparatory motor activity associated with the forthcoming voluntary motor act. However, it is not clear whether this neural signature is associated with preparatory motor activity, expectation of its associated sensory consequences, or both. Here we recorded EEG data from 14 healthy subjects while they performed self-paced button presses with their right index and middle fingers. Button-presses with one finger triggered a sound (motor+sound condition), while button-presses with the other finger did not (motor-only condition). Additionally, subjects listened to externally-generated sounds delivered in expected timings (sound-only condition). We found that the RP amplitude (locked to time of button press) was significantly more negative in the motor+sound compared with motor-only conditions. Importantly, no signal negativity was observed prior to expected sound delivery in the sound-only condition. Thus, the differences in RP amplitude between motor+sound and motor-only conditions are beyond differences in mere expectation of a forthcoming auditory sound. Our results suggest that information regarding expected auditory consequences is represented in the RP preceding voluntary action execution.

Introduction

Self-generated, voluntary actions are rarely performed without a preceding preparation or planning period during which various decisions regarding the upcoming actions’ time, trajectory and goal are made. Accumulating research over the past decades suggests that neural activity during the time interval preceding voluntary action execution is associated with different aspects of the forthcoming motor act, such as task, action type, and selection of the appropriate effector (Haggard, 2008). This view is supported by functional magnetic resonance imaging (fMRI) and electrophysiological studies in humans and primates showing that the neural activity during preparatory time intervals preceding action execution can represent, for example, the type of actions (grasping or touching; Gallivan et al., 2011), the executing effector (right or left hand; Soon et al., 2008) or the tool that is about to be used (Brandi et al., 2014; see also Bulea et al., 2014; Cisek and Kalaska, 2004; Perez et al., 2015)

Self-initiated, voluntary actions are usually preceded by a slow negativity in the scalp electroencephalography (EEG) recorded from frontal and central regions (termed “readiness potential”, RP; Kornhuber and Deecke, 1990; Libet et al., 1983). This negativity, is usually divided into early and late phases, starting around ~1500 ms and ~500 ms prior to action execution, respectively (Shibasaki and Hallett, 2006). The early phase is believed to reflect gradual increase in neural firing rate in high motor cortical regions such as supplementary motor area (SMA) and are associated with preparation and initiation of the forthcoming motor act (Lang et al., 1991, Yazawa et al., 2000, Fried et al., 2011, Pedersen et al., 1998, Cunnington et al., 2003). The late phase is more lateralized and specific to the motor command of the executing effector and is therefore believed to represent preparatory activity in primary motor cortex (M1; Passingham, 1987; Pedersen et al., 1998).

When an unimanual voluntary action is performed, for example, by a right hand, stronger negativity is usually found over the left hemisphere - contralateral to the active hand (Kutas and Donchin, 1980). This difference in RP has been shown to begin in close temporal proximity to initiation of the motor act (De Jong et al., 1988, Gratton et al., 1988, Smid et al., 1987) and is usually detected by subtracting the RP signal over the ipsilateral hemisphere from that of the contralateral hemisphere with respect to the active hand. This late RP component, termed the lateralized readiness potential (LRP), is interpreted as manifestation of preparatory brain activity that is more specific to the executing effector which is about to perform the action.

Despite recent evidence suggesting that RP does not necessarily reflect motor preparation, but might be associated with general decision making processes (Alexander et al., 2016), this neural signature is predominantly discussed in the framework of a neural correlate of motor intention and consciousness of such intention (Haggard and Eimer, 1999, Libet et al., 1983). Nevertheless, the link between RP and self-generated sensory consequences can be formulated within the forward model of sensory-motor integration (Wolpert et al., 1995). According to the model, during the preparatory period prior to voluntary action execution, the motor cortex sends signals both to the executing effectors and to the relevant sensory cortex, in which the sensory consequences of the action are expected. The signal that is sent from motor to sensory cortices was termed “efference copy” (von Holst, 1954) and was proposed to facilitate, attenuate, or otherwise modulate both perceptual and neural responses to self-generated stimuli compared with responses to identical stimuli generated by an external source (Crapse and Sommer, 2008). Therefore, the RP and its LRP subcomponent, are candidate neurophysiological signatures for such a forward model embedding the expected sensory consequences of self-generated actions.

The aim of the current study was to examine whether the RP and/or LRP encode the expected sensory consequences triggered by voluntary actions. To this end, we recorded EEG data from healthy subjects while they performed voluntary, self-paced button presses either with or without expected auditory consequences.