Coordinates for the somatosensory homunculus

In the text surrounding their drawing of the simiusculus (little monkey), Clinton Woolsey and colleagues (Woolsey et al. 1952) wrote “It must be emphasized, however, that this diagram is an inadequate representation of the localization pattern, since in a line drawing one cannot indicate the successive overlap which is so characteristic a feature of cortical representation, not only in the motor but also in the sensory areas.” Yet when most people, including neurophysiologists, look at a simiusculus or homunculus (little human) they often come away with an initial impression of a point-to-point two-dimensional map, like a map of the surface of the Earth. The study by Roux and colleagues in the current issue of The Journal of Physiology (Roux et al. 2018), by utilizing a standardized coordinate frame, provides a fresh perspective on details of somatotopic organization in the human primary somatosensory cortex (S1), including valuable details that are difficult to portray in a two-dimensional drawing. Similar to Penfield and his colleagues 80 years ago (Penfield & Boldrey, 1937), Roux and colleagues compiled data on the somatic location of sensations evoked by intraoperative electrical stimulation along the crown of the human postcentral gyrus. Such data were collated from 50 patients with intact somatic sensation who were undergoing open neurosurgical procedures for lesions that did not involve the postcentral gyrus. But whereas Penfield simply marked the stimulated locations on a standard drawing of the human cerebral cortex, Roux and colleagues recorded the 3-dimensional coordinates of each location with a neuro-navigation system and subsequently normalized those patient-specific coordinates into the standardized Montreal Neurological Institute (MNI) coordinate frame. (The Institute was founded by Penfield in 1934; the coordinate system was developed in the 1990s based on magnetic resonance images, not available in Penfield's time.) This more quantitative approach revealed novel aspects of somatotopic organization in human S1. The authors ultimately plotted the somatic location of the evoked sensation as a function of the MNI coordinates of their simulating electrode, focusing on the mediolateral dimension (their Fig. 5; Roux et al. 2018). While confirming the mediolateral sequence described by Penfield and colleagues, this advanced approach revealed the overlap intimated by Woolsey based on his studies in non-human primates. The locations at which the thumb, index, middle, ring and little fingers are represented, though sequential on average, also overlap considerably, not only with one another but also with the locations at which the palm, wrist, and forearm are represented. Similar overlap is evident in the locations at which the lips, jaw and tongue are represented. Attempting to show such overlap in a drawing of the homunculus would produce a nebulous illustration. Moreover, the MNI coordinates can be subjected to quantitative comparisons. Figures 2A and 5 of Roux et al. (2018), for example, suggest that the representations of the five digits, while occurring in the same somatotopic sequence in the left and right hemispheres, might be more dispersed and therefore overlapping in the right hemisphere than in the left. Use of the standardized MNI coordinates enabled the authors to show that this difference in the dispersion of individual digit representations between the left and the right S1 is statistically significant. As the authors point out, the number of loci that can be stimulated intraoperatively in any given human patient is necessarily limited by concerns for the clinical goals of the procedure and the patient's safety. One may be tempted to think, therefore, that two features – (i) the dispersion in the location of representation of a given digit, and (ii) the overlap of representations of digits nearby – both result from inter-individual variation of an underlying representation that has little dispersion and minimal overlap in any single subject. To address this possibility, studies are needed in which the somatotopic organization of human S1 can be investigated more extensively in individual patients using physiological techniques. Current efforts directed at the development of neuroprosthetic devices are providing such opportunities via electrocorticography grids (Hiremath et al. 2017) and penetrating microelectrode arrays (Flesher et al. 2016) implanted in human patients for days to months. Interestingly, such studies have noted that stimulation not uncommonly produces a sensation on multiple digits, whereas Roux et al. report such a diffuse sensation from only 2 of 258 responsive sites, possibly reflecting differences in technical details, such as monopolar versus bipolar stimulation or supra-threshold versus threshold stimulation. Electrical stimulation of the brain is inherently unnatural. Numerous neuron somata, axons and dendrites in the vicinity of the electrode(s) are all stimulated synchronously, in contrast to their natural activity which is by and large asynchronous at the millisecond time scale. In experimental animals, physiologists therefore have examined neural responses in S1 to touching the skin surface, passively rotating joints, and/or squeezing muscles, using both evoked potentials and evoked unit discharges. These more naturalistic approaches to investigating the organization of S1 eventually may be undertaken in humans using implanted electrode arrays. Such studies in non-human primates have shown that almost 40% of neurons in Brodmann's area 1, the area typically stimulated in humans, have multi-digit receptive fields, both cutaneous and deep (Iwamura et al. 1983). Neurons with receptive fields confined to single digits are more prevalent in area 3b. However, unlike area 1 on the surface of the postcentral gyrus, the human area 3b lies in the posterior bank of the central sulcus, where to date it remains relatively inaccessible to focal electrical stimulation. Future studies that utilize combinations of standardized coordinates, implanted electrode arrays, and naturalistic activation are likely to provide an even more detailed understanding of somatotopic organization in human S1. None declared. This work was supported in part by R01 NS079664 and R01 NS092626 from the NINDS. The author thanks Marsha Hayles for editorial comments.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.