Combining correlation and interference methods in the human brain. Focus on "Cortico-cortical interactions in spatial attention: A combined ERP/TMS study".

Recent years have witnessed an advance in the understanding and measurement of the neural correlates of attention in the human brain. For example, it is established that the allocation of attention is accompanied by changes in event related electrical potentials (ERPs) that can be recorded at the scalp. In addition, it is known that visuospatial attention is disrupted when transcranial magnetic stimulation (TMS) is applied over right posterior parietal cortex (rPPC). It has been unclear, however, whether the effect of the TMS is mediated by the same mechanisms that are indexed by the ERP. In this issue of J. Neurophysiology (p. 3277–3280), Fuggetta et al. (2006) have managed to record ERPs while they simultaneously applied TMS to the rPPC. They report that the ERP signature of attention is disrupted by the application of TMS. Specifically, rPPC TMS eliminates an ERP component called the N2pc. The possibility of recording ERPs while simultaneously applying TMS has excited many human cognitive neuroscientists because it raises the possibility of directly examining, in the human brain, the causal impact of activity in one area on activity in another area. There have been a few reports of electroencephalographic activity (EEG) changes in response to TMS (Ilmoniemi et al. 1997; Paus et al. 2001), but the present study is one of the first to demonstrate that TMS can have an impact on an ERP that is associated with a particular cognitive process. Fuggetta’s technically innovative study builds on previous experiments that employed either just TMS or ERP recording during visual search tasks. In such tasks, subjects are shown an array of visual stimuli and decide whether a certain target is present. The decision is difficult and attentionally demanding when distractor stimuli, which are visually similar to the target, are also presented. Behavioral measurements show that application of TMS over the rPPC just after presentation of the search array disrupts target identification (Ashbridge et al. 1997; Rosenthal et al. 2005). The effect is especially strong on “conjunction” search for targets defined by a combination of features as opposed to when a single feature makes the target ‘pop-out’ (Ellison et al. 2004). The N2pc is a negative ERP component elicited by search arrays containing the target stimulus that can be recorded 200 ms after onset of the visual stimulus array. It is distributed over the posterior electrodes contralateral to the target in the array. Conjunction search elicits a stronger N2pc than pop-out search (Luck and Hillyard 1994). By combining TMS and ERP techniques, it is possible to test whether rPPC activity is a critical determinant of the N2pc. There is a methodological challenge to be overcome in simultaneously applying TMS and recording EEG: an unchecked TMS pulse saturates conventional EEG amplifiers. This problem was first solved by Ilmoniemi et al. (1997) by using a preamplifier with externally gated microswitches to transiently disconnect the EEG amplifier from the EEG electrodes for a brief time window around the TMS pulse. After switching there can be a brief artifact as the amplifier resets its baseline. Fuggetta et al. employed a different method, using one of a new breed of amplifiers that have wide operating ranges and so do not get saturated by the brief but strong TMS pulse. The brief spike in the EEG caused by the pulse is excised from the raw EEG trace so as not to interact with off-line filtering of the signal. To control for the neural activity reflecting processing of the acoustic and somatosensory stimuli accompanying the TMS pulse, raw EEG data from each trial is cleaned by subtracting from it a “template” ERP waveform. The template is derived by recording the average ERP elicited by a TMS pulse in the absence of a task. In practice, this approach can also result in a brief loss of data immediately after the TMS pulse (Thut et al. 2005). It may be the case then that the technique will not allow the investigations of interareal interactions at quite the same time scale as is possible in animal studies (Moore and Armstrong 2003). Nevertheless, the technique is still likely to prove useful in studies of the human brain because the synchronized neural activity that constitutes the ERP signal may take some time to develop. Certainly in the present study the application of TMS 100 ms after stimulus onset had an effect on an ERP that was recorded a further 150 ms later. The Fuggetta study is no doubt just one of the first of many to examine the effect of TMS on electrophysiological indices of cognitive processes. One of the intriguing features of the N2pc in the current study is that the location of its source in the brain is not clear. Hopf et al. (2000) have used magnetoencephalographic (MEG) data to argue that it originates in both parietal and ventral occipitotemporal areas. Whether the rPPC stimulation in the present study is disrupting the N2pc that is generated in the parietal cortex directly under the TMS coil or whether it also affects the N2pc that is generated at some distance away in occipitotemporal cortex is not clear. An unambiguous demonstration of the effect of TMS of one area on an ERP known to be generated by another area would open the door to the investigation of interactions between brain areas during cognition. With the TMS-ERP combination, it might then Address for reprint requests and other correspondence: P.C.J. Taylor, Dept. of Experimental Psychology, South Parks Rd., Oxford OX1 3UD UK (E-mail: [email protected]). J Neurophysiol 95: 2731–2732, 2006. doi:10.1152/jn.00058.2006.