Research report
The pre-movement component of motor cortical local field potentials reflects the level of expectancy

https://doi.org/10.1016/j.bbr.2006.02.004Get rights and content

Abstract

Cortical local field potentials (LFPs) are modulated in parallel with single neuron discharge, but the information they carry is often unclear. Multi-electrode recordings of both LFPs and single neuron activities were made in motor cortex as monkeys performed a delayed pointing task in which the probability of the moment of signal occurrence, and thus movement execution, was manipulated. A large positive LFP component (P1) appeared immediately preceding movement onset only under conditions of low probability, that is, when a response signal was weakly expected. The amplitude of P1 was much smaller when probability of signal occurrence was high, or when the same movement was self-paced. Although P1 has been described as being linked to the descending motor signal, we found that it was more closely associated with the processing of movement-related information than with the ultimate motor command. Its timing did not bear a fixed relationship with movement onset and its frequency of occurrence in each monkey varied in parallel with each animal's overall performance and the percentage of context-related “pre-processing” neurons encountered.

Introduction

Many features of motor cortical single neuron activity are not strictly related to the motor signal, i.e. movement generation. Instead they are more closely tied to preparation for a planned movement, or even suppression of movement until the appropriate moment (for a review, see [25]). For example, some activity has been described as “pre-processing” [28], i.e. processing of prior information about features of the forthcoming movement. Furthermore, the transient synchronization of two or more neurons at the end of an estimated time interval can reflect instants of signal expectancy [26], [27]. Such properties are linked to the temporal prediction of forthcoming events, which is essential for optimizing motor performance. Indeed, prior information about when a response signal will occur significantly shortens reaction time [22], [25]. If these processes arise in a sufficiently large population of neurons, they should also be manifested in the local field potential (LFP), possibly indicating the degree of coherent network activity. LFPs may be recorded from the same electrode as the spikes of a single neuron, by simply low-pass filtering the signal. The LFP is a spatially averaged signal from a small volume, assumed to reflect mainly the synaptic input received by the neurons within the observed volume [17], [18].

Studies of LFPs recorded in motor cortical areas have mainly focused on three topics: firstly, stimulus-induced (but not stimulus-locked) oscillatory activity ([2], [4], [19], [20], [21], [32]; for a review, see [13], [14]), secondly, movement-aligned potentials [7], [24], and thirdly, correlations between LFPs recorded on different electrodes, including those from different hemispheres [3]. To our knowledge, Gemba and co-workers were the first to study monkey cortical field potentials in relation to both visually-triggered and self-paced movements. They identified a depth-positive (surface-negative) potential, immediately preceding movement onset [12], that was much larger in motor cortex than in adjacent regions [10]. The pre-movement positivity was interpreted as being strongly motor-related; it disappeared when an operator performed the task, instead of the monkey executing the movement [10]. Since then, the pre-movement positivity has been observed by others (e.g. [7], [16], [24]), and is generally assumed to be linked to the generation of the descending motor volley. It is not, however, significantly modulated by movement direction, although later LFP components are directionally sensitive [24].

We recorded LFPs in two monkeys performing two tasks, a choice reaction time task (chRT) and a self-paced movement task (SELF), first presented in ref. [29]. In both tasks, accurate estimation of two time intervals was essential for correct performance. Surprisingly, we observed a large context-related modulation of the pre-movement positivity (P1). When aligned to movement onset, this pre-movement positive component exhibited amplitude differences in relation to the probability of the moment of the response signal, and timing differences in relation to reaction time. Moreover, the presence of P1 depended strongly on the overall behavioral performance of the monkey. Later LFP components (after movement onset) did not exhibit these properties. Preliminary results were presented in ref. [30].

Section snippets

Experimental procedures

Behavioral tasks (Fig. 1, see [29]): Three male Rhesus monkeys were used in this study. Care and treatment of the animals during all stages of the experiments conformed to the European and French Government Regulations. The animals were trained to execute movements in two opposite directions from a common center position. LFPs were recorded in only two of them, monkeys O and K; the third, monkey R, reported in ref. [29], is included for comparison. On a vertical panel, three touch-sensitive

Behavioral results and EMG data

Learning to discriminate or to estimate two time intervals took several months of daily training. Once the monkey learned the durations, the number of errors decreased and the performance, in terms of reaction time, improved. Performance varied largely from monkey to monkey (for the third monkey trained in this task, monkey R, see [29]). Distributions of trial-by-trial response times are shown in Fig. 1C–F for monkeys K and O, in both tasks recorded during all selected sessions (for selection

The pre-movement positivity: an information processing LFP component

A pre-movement positivity (depth-positive–surface-negative), similar to our P1 component, was described by Gemba et al. [10] in cortical field potentials and interpreted as a motor component. We have shown that P1 has properties that are inconsistent with any direct relationship to the descending motor volley. For the same movement, P1 can be very large or virtually non-existent, depending on the degree of uncertainty about the time of RS. In the present data, the P1 component attained its

Acknowledgements

We thank Driss Boussaoud, Bjørg Kilavik, Martin Nawrot, Alex Thomson and Eilon Vaadia for valuable discussions of an earlier version of the manuscript and Martin Nawrot for help in data analysis. WAM was supported by the CNRS (“poste rouge”) during a sabbatical. The research was supported in part by the CNRS and the French Government (ACI Cognitique “Variability and Invariants”).

References (33)

  • R. Courtemanche et al.

    Local field potential oscillations in cerebellar cortex: synchronization with cerebral cortex during active and passive expectancy

    J Neurophysiol

    (2005)
  • D.J. Crammond et al.

    Prior information in motor and premotor cortex: activity during the delay period and effect on pre-movement activity

    J Neurophysiol

    (2000)
  • A. Destexhe et al.

    Spatiotemporal analysis of local field potentials and unit discharges in cat cerebral cortex during natural wake and sleep phases

    J Neurosci

    (1999)
  • O. Donchin et al.

    Local field potentials related to bimanual movements in the primary and supplementary motor cortices

    Exp Brain Res

    (2001)
  • J.J. Eggermont et al.

    Azimuth coding in primary auditory cortex of the cat. I. Spike synchrony versus spike count representations

    J Neurophysiol

    (1998)
  • E.V. Evarts

    Pyramidal tract activity associated with a conditioned hand movement in the monkey

    J Neurophysiol

    (1966)
  • Cited by (32)

    • Modeling the spatial reach of the LFP

      2011, Neuron
      Citation Excerpt :

      The local field potential (LFP) usually refers to the low-frequency part (≲ 500 Hz) of an extracellular voltage signal recorded inside the brain. It is among the oldest experimental measures of neural activity and has been widely used to investigate network mechanisms involved in sensory processing (Mitzdorf, 1985; Di et al., 1990; Kandel and Buzsáki, 1997; Schroeder et al., 1998; Henrie and Shapley, 2005; Belitski et al., 2008; Montemurro et al., 2008; Szymanski et al., 2009), motor planning (Scherberger et al., 2005; Roux et al., 2006), and higher cognitive processes including attention, memory, and perception (Pesaran et al., 2002; Kreiman et al., 2006; Liu and Newsome, 2006; Womelsdorf et al., 2006; Montgomery and Buzsáki, 2007; Colgin et al., 2009). In combination with multiunit activity (MUA), the high-frequency (≳ 500 Hz) part of the extracellular voltage, it has been found useful for inferring key properties of network dynamics (Denker et al., 2010, 2011; Kelly et al., 2010) and population-specific laminar activity (Einevoll et al., 2007).

    • Pre-stimulus beta oscillations within left posterior sylvian regions impact auditory temporal order judgment accuracy

      2011, International Journal of Psychophysiology
      Citation Excerpt :

      Such processes have been demonstrated in a study by Van Ede et al. (2010), showing that pre-stimulus beta power modulates depending on participants' preparation of the arrival of expected somatosensory stimuli, more beta power, however, decreasing performance. Baseline beta fluctuation might have followed from changes in alertness (Roux et al., 2006; Buschman and Miller, 2007, 2009), which in turn have been shown to modulate gating processes (Guterman and Josiassen, 1994; Guterman, et al., 1992). This hypothesis is compatible with our result for slower RTs in the inaccurate than accurate condition.

    • Fast and Slow Oscillations in Human Primary Motor Cortex Predict Oncoming Behaviorally Relevant Cues

      2010, Neuron
      Citation Excerpt :

      Experimental findings have lead to two common interpretations of the functional relevance of beta oscillations. The first is that they are related to maintaining a stable posture by inhibiting movement (Baker et al., 1999; Kühn et al., 2008b), and the second that they are related to some aspect of movement planning or attention (Donoghue et al., 1998; Murthy and Fetz, 1992, 1996; Roux et al., 2006; Sanes and Donoghue, 1993; Schwartz et al., 2005). Previous reports showed that the beta power increases during an instructed delay period (Donoghue et al., 1998; O'Leary and Hatsopoulos, 2006) but did not test for the task-relevance of the instruction.

    • Influence of Uncertainty and Surprise on Human Corticospinal Excitability during Preparation for Action

      2008, Current Biology
      Citation Excerpt :

      Learning the relative probabilities of impending actions may enable the nervous system to prepare motor output prior to an event. Sensory cues that predict action enable a gradual build-up of preparatory activity in premotor and motor cortex prior to action [4–10, 20]. This build-up is reflected by specific excitability changes in corticospinal projections [15, 16].

    View all citing articles on Scopus
    1

    Tel.: +1 416 978 2675; fax: +1 416 978 4940.

    View full text