In the back of your mind: Cortical mapping of paraspinal afferent inputs

Abstract

Persistent pain alters brain-body representations, highlighting their potential pathological significance. In chronic low back pain (LBP), sparse evidence points towards a shift of the cortical representation of sensory afferents of the back. However, systematic investigations of the cortical representation of tactile and proprioceptive paraspinal afferents along the thoracolumbar axis are lacking. Detailed cortical maps of paraspinal afferent input might be crucial to further explore potential relationships between brain changes and the development and maintenance of chronic LBP. We therefore validated a novel and functional magnetic resonance imaging- (fMRI-)compatible method of mapping cortical representations of tactile and proprioceptive afferents of the back, using pneumatic vibrotactile stimulation ("pneuVID") at varying frequencies and paraspinal locations, in conjunction with high-resolution fMRI. We hypothesised that: (i) high (80 Hz) frequency stimulation would lead to increased postural sway compared to low (20 Hz) stimulation, due to differential evoked mechanoreceptor contributions to postural control (proprioceptive vs tactile); and (ii) that high (80 Hz) versus low (20 Hz) frequency stimulation would be associated with neuronal activity in distinct primary somatosensory (S1) and motor (M1) cortical targets of tactile and proprioceptive afferents (N=15, healthy volunteers). Additionally, we expected neural representations to vary spatially along the thoracolumbar axis. We found significant differences between neural representations of low and high frequency stimulation and between representations of thoracic and lumbar paraspinal locations, in several bilateral sensorimotor cortical regions. Proprioceptive (80 Hz) stimulation preferentially activated sub-regions S1 3a and M1 4p, while tactile (20 Hz) stimulation was more encoded in S1 3b and M1 4a. Moreover, in S1, lower back proprioceptive stimulation activated dorsal-posterior representations, compared to ventral-anterior representations activated by upper back stimulation. As per our hypotheses, we found distinct sensorimotor cortical tactile and proprioceptive representations, with the latter displaying clear topographic differences between the upper and lower back. This thus represents the first behavioural and neurobiological validation of the novel pneuVID method for stimulating muscle spindles and mapping cortical representations of paraspinal afferents. Future investigations of detailed cortical maps will be of major importance in elucidating the role of cortical reorganization in the pathophysiology of chronic LBP.

BackWards – Unveiling the Brain’s Topographic Organization of Paraspinal Sensory Input

“…In addition, the results are in line with the general observation that the spatial relationship between receptors in the body is maintained along the body-brain axis, reflected by somatotopically organized representations in the sensory homunculus (Yamada et al, 2007, Puckett et al, 2020, Schott, 1993). Furthermore, although no bilateral representations were found in M1, the extension of the model-based representational activation cluster into BA4p of the right hemisphere indicates that the central processing of paraspinal vibrotactile afferent input involves integrating contributions from somatosensory and motor cortices as previously suggested (Asanuma et al, 1980, Avanzino et al, 2014, Cole et al, 2022).…”

“…Full details on the pneuVID methodology and its integration into the MRI environment have previously been outlined (Schibli et al, 2021, Cole et al, 2022). In short: vibrotactile stimulation units are attached bilaterally to the identified spinous processes; a valve box delivers pulses of compressed air to the vibrational units; and a control module (Raspberry Pi, Cambridge, UK) connects the setup to the MRI system.…”

“…The M1 mask comprises Brodmann areas (BA) BA4a and BA4p and the S1 mask BA3a, BA3b, BA1, and BA2 as well as BA5 of the posterior parietal cortex. The choice of these two masks was motivated by the previously published results on neural processing of tactile and vibrotactile stimulation (Delhaye et al, 2018, Schellekens et al, 2021, Cole et al, 2022). The probabilistic atlas threshold was set at 20% causing some overlap between the two masks.…”

“…Pre-processing of acquired neuroimaging data followed a previously published pipeline (Cole et al, 2022).…”

“…Paradigms using vibrotactile stimuli, which correspond to a naturalistic afferent input, are an established way of assessing cortical somatotopic representations from different body parts (Harrington and Downs III, 2001, Kim et al, 2016, Schellekens et al, 2021). The usefulness of pneuVID to assess the neural representations of the human back has recently been validated using univariate statistical neuroimaging techniques (Cole et al, 2022). Representations based on univariate analysis are a first step towards establishing the cortical mapping of the healthy human back.…”