Pre- and post-stimulus alpha activity shows differential modulation with spatial attention during the processing of pain

Abstract

Extensive work using magneto- and electroencephalography (M/EEG) suggests that cortical alpha activity represents a top-down controlled gating mechanism employed by processes like attention across different modalities. However, it is not yet clear to what extent this presumed gating function of alpha activity also applies to the processing of pain.

In the current study, a spatial attention paradigm was employed requiring subjects to attend to painful laser stimuli on one hand while ignoring stimuli on the other hand. Simultaneously, brain activity was recorded with MEG. In order to disentangle pre- and post-stimulus effects of attention, alpha activity was analyzed during time windows in anticipation of and in response to painful laser stimulation.

Painful laser stimuli led to a suppression of alpha activity over both ipsi- and contralateral primary somatosensory areas irrespective if they were attended or ignored. Spatial attention was associated with a lateralization of anticipatory pre-stimulus alpha activity. Alpha activity was lower over primary somatosensory areas when the contralateral hand was attended compared to when the ipsilateral hand was attended, in line with the notion that oscillatory alpha activity regulates the flow of incoming information by engaging and/or disengaging early sensory areas. On the contrary, post-stimulus alpha activity, for stimuli on either hand, was consistently decreased with attention over contralateral areas. Most likely, this finding reflects an increased cortical activation and enhanced alerting if a painful stimulus is attended.

The present results show that spatial attention results in a modulation of both pre- and post-stimulus alpha activity associated with pain. This flexible regulation of alpha activity matches findings from other modalities. We conclude that the assumed functional role of alpha activity as a top-down controlled gating mechanism includes pain processing and most likely represents a unified mechanism used throughout the brain.

Introduction

Rhythmic neuronal alpha oscillations between 8 and 12 Hz are the strongest electrophysiological signal measured from the awake human brain (Berger, 1929, Niedermeyer and da Silva, 2005). For a long time, alpha activity was thought to merely reflect “cortical idling” (Pfurtscheller et al., 1996). Recently, alpha activity has been assumed to have a more active role in terms of a graded functional inhibition of brain areas not directly involved in a specific task (Jensen and Mazaheri, 2010). In this context, alpha activity has been closely related to selective attention and is understood to actively gate the incoming flow of information (Foxe and Snyder, 2011), enabling us to favor and more efficiently process those of the many incoming stimuli in our environment that are behaviorally relevant. Across different modalities, attention affects oscillatory alpha activity already during anticipatory pre-stimulus periods, i.e. when a stimulus is expected but not yet actually received and processed (Del Percio et al., 2006, Jones et al., 2010, Thorpe et al., 2012, Thut et al., 2006). In the somatosensory system, for example, spatial attention to one hand lateralizes alpha activity in anticipation of a tactile stimulus. Specifically, alpha activity is decreased in the primary somatosensory cortex contralateral to the attended hand, but increased ipsilaterally (Anderson and Ding, 2011, Haegens et al., 2011). Moreover, it has been shown that the degree of pre-stimulus alpha lateralization determines the subsequent behavioral response (Haegens et al., 2011). These findings tally the notion that an increase of alpha activity reflects an active inhibition or disengagement of a cortical area (Foxe and Snyder, 2011, Jensen and Mazaheri, 2010) whereas a decrease of alpha activity is a correlate of an activated or engaged cortical region (Pfurtscheller et al., 1996).

Painful stimuli have been demonstrated to affect oscillatory activity across a range of frequency bands (Hauck et al., 2007, Hauck et al., 2008, Schulz et al., 2012). With respect to lower frequencies, previous work showed that painful laser stimuli, in contrast to stimuli of other modalities (Hari and Salmelin, 1997), globally suppress spontaneous alpha and beta oscillations in somatosensory, motor and visual areas (Ploner et al., 2006a). This finding has been interpreted as a specific alerting function of pain, which is also reflected by an increased excitability of somatosensory cortices to subsequent tactile stimuli (Ploner et al., 2006b). Interestingly, seeing an image of a limb in a painful situation also induces stronger alpha suppression in sensorimotor areas than seeing an image of a non-painful situation (Whitmarsh et al., 2011). Together, these results suggest that the processing of pain and pain-associated stimuli in somatosensory cortices is particularly intense and associated with a strong modulation of oscillatory activity.

The relationship between pain-associated oscillatory alpha activity and attention is not yet fully understood. In a study using subdural electrocorticographic recordings (ECoG) from epilepsy patients, Ohara et al. (2004) found that alpha suppression in response to an attended compared to a non-attended painful laser stimulus is more intense and widespread over primary somatosensory and parasylvian cortices. In contrast, using an oddball paradigm and painful intracutaneous electrical stimulation in healthy subjects, Hauck et al. (2007) did not observe any modulation of pain-associated alpha activity with attention. Both studies, however, did not investigate possible pre-stimulus effects of attention. Previous work addressing pre-stimulus effects showed that anticipation of a painful stimulus already results in a suppression of alpha activity (Babiloni et al., 2003, Babiloni et al., 2004, Babiloni et al., 2006, Del Percio et al., 2006). This anticipatory suppression is again stronger for a painful compared to a non-painful stimulus (Babiloni et al., 2003) and less prominent, if the subject is distracted by mentally performing an arithmetical task (Del Percio et al., 2006).

Linking the attentional modulation of pain processing to behavior, previous studies have consistently shown that a stimulus is perceived as less painful if attention is directed away from it (Bushnell et al., 1999, Miron et al., 1989, Petrovic et al., 2000, Schlereth et al., 2003). Yet, even under constant experimental conditions, the pain experience varies substantially both across and within subjects (Babiloni et al., 2006, Gross et al., 2007, Schulz et al., 2011, Schulz et al., 2012). These variations covary with alpha activity. For example, the strength of anticipatory pre-stimulus alpha suppression in expectation of a painful stimulus correlates with the intensity of the individual pain perception across subjects (Babiloni et al., 2006). Within subjects, pain ratings across different trials with stimuli of constant intensity vary in close relation to the post-stimulus pain-induced alpha suppression (Schulz et al., 2011). Therefore, the pain sensation seems to be related to the current state of alpha activity. More specifically, pain appears to be perceived more intensely, if alpha activity close to the moment of stimulus onset is low.

The current study intended to shed further light on the role of alpha activity in relation to pain processing. Previous studies indicate both a modulation of pain-related alpha activity by attention and an association between alpha activity and the subjective pain experience. However, concurrent effects of attention on both pre- and post-stimulus pain-associated alpha activity have not yet been addressed. We aimed to test the hypothesis that alpha activity has a similar function for pain processing as for other modalities, representing a gating mechanism employed by top-down processes to streamline information flow in the brain. To this end, we used a spatial attention paradigm requiring subjects to attend to painful laser stimuli on the dorsum of one hand, while at the same time ignoring stimuli on the other hand. Simultaneously, brain activity was recorded using magnetoencephalography (MEG). To disentangle pre- and post-stimulus attention effects, we analyzed alpha activity both in anticipation of and in response to painful laser stimuli. For the anticipatory pre-stimulus period, spatial attention was hypothesized to lateralize alpha activity across left and right somatosensory cortices. In line with an active regulation of the incoming information flow, alpha activity was expected to be lower in primary somatosensory areas when the contralateral hand was attended compared to when the ipsilateral hand was attended. For the post-stimulus period, attention was assumed to decrease alpha activity primarily over contralateral somatosensory areas, in accordance with an increased cortical activation. Lastly, we expected attention-related alpha activity to be predictive of the individual pain perception.