On the challenge of measuring direct cortical reactivity by TMS-EEG.

Since its introduction in 1997 by Ilmoniemi and colleagues [1], the combination of TMS and EEG has been proposed as a new and unique method of characterizing brain reactivity and connectivity [2]. However, shortly after its introduction it became clear that this combination brings new technical problems. At the beginning, researchers’ attentionwas focused on the saturation of the EEG amplifiers caused by the TMS pulse, which has led to the introduction of several TMS-compatible recording systems (for a more extensive description see Ref. [3]) and to the definition of the ideal recordings parameters with the attempt to maximally reduce the artifact duration [3,4]. Beside recording settings, several off-line procedures have been proposed to solve the artifact issue such as subtraction approaches, filtering methods, principal component analysis (PCA) and independent component analysis (ICA). Unfortunately, these methods are not established standard procedures to remove themagnetic artifact and thus have not been further applied. Therefore, the best possibility so far to reduce the artifact duration is selecting the right recording parameters, with independent groups reporting a reasonable loss of EEG signal, that is 5e6 ms from the pulse delivery [3,4]. More recently, researchers have focused on the EEG signal recorded after the magnetic artifact, namely the TMS-evoked potentials (TEPs) recorded in the first milliseconds following the pulse delivery (in the first 5e10 ms following pulse), as spurious extracortical sources may contribute to the generation of these components. The debate about the short latency TEPs startedwhen several groups reported a huge bipolar waveform recorded between 5 and 10 ms (P5, N8) after TMS pulse [4e6], mainly evoked when the stimulation was performed over lateral scalp positions. Despite the source of this waveform is unknown, it has been interpreted as muscle activity [5], huge cortical response or as a combination of cortical and extra-cortical signals [6,7]. The first conclusion is mainly supported by a recent work by Mutanen et al. [5], nicely showing that the amplitude of the biphasic response is highly dependent on coil distance, orientation and tilt angle relative to cranial muscles. In particular, its amplitude decreased when the coil, or even its wings, was moved toward central sites and it was no longer recorded when TMS was applied over the midline. To further explore the nature of these early responses, Rogasch et al. [4] and Veniero et al. [6,7] tried to manipulate its amplitude by means of TMS protocols known to modulate cortical excitability. However, Rogasch et al. [4] found no modulation of early components following inhibitory paired-pulse technique, whereas Veniero et al. [6,7] found a modulation of P5 and N8 components after applying rTMS over primary motor cortex or premotor area.