Introduction

Studies in monkeys indicate that somatosensory processing of innocuous tactile stimuli occurs within an interconnected cortical network, which mostly resides in the parietal cortex. The major regions include multiple subdivisions of the primary somatosensory area (SI) in the postcentral gyrus, the secondary somatosensory area (SII) in the parietal operculum, additional lateral cortical areas buried within the Sylvian fissure, portions of the supramarginal gyrus, and granular prefrontal cortex. This review focuses on homologous somatosensory representations in the cortex of humans that are revealed with non-invasive recording and neuroimaging methods. Critical to the validation of these observations is confirming characteristic features known from direct neural recordings, especially in monkeys. Brain imaging studies like positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) are especially appropriate in exposing the many cortical foci that respond during a single somatosensory stimulation paradigm. Non-invasive techniques have the great advantage of allowing replicate studies in conscious normal or patient populations performing a variety of tasks affecting tactile discrimination, sensorimotor integration, attention, and object recognition.

A discussion of different brain imaging techniques is beyond the scope of this review. However, some comments about underlying characteristics of these images are useful in evaluating the findings on somatosensory areas. PET and fMRI images are a consequence of local blood flow changes, which accompany immediate and focal neuronal activity. These images have low signal-to-noise ratios, are delayed in seconds from the millisecond intervals of neural activity by the slowness of hemodynamic responses, and are spatially constrained to the vascular distribution. However limited is the spatial resolution of PET or fMRI, the localization of activated areas is direct.

Magnetoencephalography (MEG) records evoked neural responses much like electroencephalography and, therefore, contribute explicit information about timing of neural activity. Signals have high reliability as they are based on averages of multiple, synchronized trials. Localizing the source of recorded electromagnetic fields requires inverse modeling to locate and orient the source and magnitude of responsible primary currents.128 Although there are infinite solutions to the inverse problem, most recorded field patterns are dipolar, which suggests that source localizations can be confined to a single equivalent current dipole (ECD). Explicitly restricting the results to MR anatomical or functional foci in single subjects improves source localization models of these ECDs. However, modeled source localizations are always problematic, especially for long latency responses, in the presence of multiple, overlapping generators, and where stronger signals possibly mask weaker responses.128

0 0

Post a comment