Motor cognition: TMS studies of action generation

The discovery and earliest applications of TMS both involved the motor system (Barker et al. 1985). Since then, TMS has been used in three quite different ways to study motor cognition. First, TMS can be used to provide a controllable and physiologically-specified input to the skeletomotor system. Several sensory studies, for example, have used TMS to generate muscle contractions in the absence of volition and movement preparation. This allows controlled psychophysical studies of the perception of bodily movement (Haggard et al. 2002; Ellaway et al. 2004; Haggard and Whitford 2004). In other studies, TMS-evoked movements are used as perturbations of the motor apparatus. Here the focus is on preparatory and reactive adjustment for the perturbation (Bonnard et al. 2003 2004). In this method, TMS is generally delivered over the primary motor cortex, but effects on the brain are less important than the effects on the body. Although this use of TMS has great value as a peripheral stimulus for studying kinesthesis, it is logically quite different from the use of TMS to study specific brain areas and processes, and so is not considered further here. A second, very important use of TMS has been as an online probe of cortical motor excitability. This is reviewed in detail elsewhere (e.g. Chapter 9, this volume). A TMS test pulse can provide a known, if artificial, input to the motor cortex. This will cause a twitch in target muscles (motor-evoked potential, MEP) whose amplitude can be precisely measured. It may also cause an inhibition of ongoing electromyogram (EMG) (silent period, SP). In cognitivemotor studies, the size of these excitatory or inhibitory effects is measured as a function of cognitive factors like task, expectancy and so forth. Changes in the motor output for a constant TMS input are interpreted in terms of differences between conditions, or across time, in motor system excitability. Importantly, this method can provide a completely implicit and on-line measure of the state of the cortical action system. Often a test pulse is preceded by a conditioning stimulus such as a sensory input or a conditioning TMS pulse to the same or another brain area. Third, TMS can be used to interfere with cognitive-motor processes involved in action control, and widely described throughout this volume. Because the brain processes involved in generating a simple action are essentially serial, a single TMS pulse delivered at an appropriate time over an actively involved brain area may disrupt action control. Such single-pulse effects tend to be highly informative, because of their temporal and spatial specificity. On the other CHAPTER 30