Combining EEG and fMRI in pain research

In 1976, Carmon et al. showed, for the first time, that radiant heat pulses generated by a CO2 laser stimulator could, when directed to the skin, elicit brain potentials in the ongoing human electroencephalogram (EEG). Such laser pulses were later demonstrated to activate Aδ and C skin nociceptors in a selective and synchronous fashion (see Plaghki and Mouraux 2003 for a review). Since this first report, numerous studies have relied on laser-evoked brain potentials (LEPs) to assess the function of nociceptive somatosensory pathways and to gain insight into the neural processes that underlie the perception of pain. In the late 1980s, a number of studies used multichannel EEG recordings to examine the topographical distribution of LEPs (Treede et al. 1988) and model their underlying neural sources, and thus started to identify the different brain areas activated by nociceptive somatosensory input (Bromm and Chen 1995; Tarkka and Treede 1993). A consistent finding across these studies is that LEPs are well explained by the combination of a midline source (usually assigned to the anterior part of the cingulate cortex, ACC), and a pair of bilateral opercular sources (usually assigned to secondary somatosensory cortex, i.e. SII, and/or insular cortex). In some studies (Tarkka and Treede 1993) an additional parietal source was added to the model and assigned to the primary somatosensory cortex (SI) contralateral to the stimulated side (see Garcia-Larrea et al. 2003 for a review). These early findings were later corroborated by a large number of studies using magnetoencephalography (MEG; Ploner et al. 2002), direct intracranial recording of local field potentials (Frot and Mauguiere 2003), and neuroimaging methods that sample neural activity indirectly by measuring stimulus-evoked changes in regional cerebral blood flow (PET and fMRI; Davis et al. 1998; Peyron et al. 1999). Several meta-analyses have reviewed the existing data on EEG, MEG, PET and fMRI responses to nociceptive stimulation (Apkarian et al. 2005; Garcia-Larrea et al. 2003; Peyron et al. 1999), and have confirmed the existence of a common set of brain regions responding to nociceptive stimuli, including bilateral thalamus, bilateral SII, bilateral insula, ACC, prefrontal cortex and, less consistently, contralateral SI cortex. A number of investigators have hypothesised that this network of brain areas, usually referred to as the “pain matrix” (Melzack 1999), reflects brain activities that are specifically involved in the processing of nociceptive input, and thus that it may constitute a “cerebral signature for pain” (Tracey and Mantyh 2007).