Our experimental design focused primarily on separately comparing S2 TMS to sham vertex TMS, and S1 TMS to sham vertex TMS. Because of the possibility that both S1 and S2 TMS are involved in pain perception, we did not have strong predictions about the differences check details between S1 and S2 conditions. Interestingly, however, we found that judgements of intensity were significantly disrupted not only when comparing S2 to vertex TMS, but also when comparing S2 to S1 TMS. This result points

to distinct roles for S1 and S2 in pain perception, even though they are co-activated in parallel (Liang et al., 2011; Ploner et al., 2009) by nociceptive stimuli. A previous study investigating the role of S1 and S2 in pain intensity discrimination observed that whilst S1 responses were able to gradually encode the intensity of a painful stimulus S2 responses had a more categorical or binary form, showing a sharp increase in amplitude at intensities above the pain threshold (Timmermann et al., 2001). Our results extend these findings by providing evidence that S2 plays a causal role in discrimination of nociceptive stimulus intensity. Kanda et al. (2003) found that TMS over S1 applied 150 msec and 200 msec post-stimulus increased reports of pain, while TMS over S2 had no effect. However,

Kanda et al.’s (2003) task focused on pain detection, rather than coding for graded levels of pain intensity. Indeed, their stimuli remained constant, and they relied on (presumably random) variations in perceived intensity. In the present study we used a two-alternative GSK1120212 molecular weight forced choice pain intensity judgement, which may be more sensitive to the neural encoding of pain levels. Our TMS did not affect participants’ ability to localise noxious stimuli. This result is consistent with the findings of Kanda et al. (2003) but at odds with those of Porro et al. (2007). These last authors observed that TMS over S1 significantly disrupted localisation of painful Acetophenone stimuli. Nevertheless, the role of S1 in pain localisation is still controversial (Apkarian et al., 2005; Bushnell et al., 1999), and several reasons could explain the discrepant results.

First, Porro et al. (2007) used mechanical stimuli that activate tactile as well as nociceptive fibres, whilst we used an Nd:YAP laser that selectively activates A-delta fibres but not A-beta fibres. The additional tactile component in Porro et al.’s (2007) study may have contributed to pain localisation, and it may have been this tactile location information that was disrupted by S1 stimulation. Further, we applied single-pulse TMS at 120 msec after a noxious stimulus, based on previous electrophysiological studies of the N1 LEP component (e.g., Valentini et al., 2012), while Porro et al. (2007) applied TMS trains 150 msec and 300 msec after a painful stimulus. They found a significant increase in localisation errors only for the later stimulation.