Distributed representation of pelvic floor muscles in human motor cortex

Abstract

Human motor cortex can activate pelvic floor muscles (PFM), but the motor cortical representation of the PFM is not well characterized. PFM representation is thought to be focused in the supplementary motor area (SMA). Here we examine the degree to which PFM representation is distributed between SMA and the primary motor cortex (M1), and how this representation is utilized to activate the PFM in different coordination patterns. We show that two types of coordination patterns involving PFM can be voluntarily accessed: one activates PFM independently of synergists and a second activates PFM prior to and in proportion with synergists (in this study, the gluteus maximus muscle - GMM). Functional magnetic resonance imaging (fMRI) showed that both coordination patterns involve overlapping activation in SMA and M1, suggesting the presence of intermingled but independent neural populations that access the different patterns. Transcranial magnetic stimulation (TMS) confirmed SMA and M1 representation for the PFM. TMS also showed that, equally for SMA and M1, PFM can be activated during rest but GMM can only be activated after voluntary drive to GMM, suggesting that these populations are distinguished by activation threshold. We conclude that PFM representation is broadly distributed in SMA and M1 in humans.

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1

Electromyographic (EMG) evidence of two muscle coordination patterns involving human pelvic floor muscles (PFM) that the brain can voluntarily access. (A) Example EMG recordings from the PFM (blue), gluteus maximus muscle (GMM, green), and the first dorsal interosseous muscle (FDI, red) in a single participant during repeated and separate voluntary activations of the PFM, GMM, and FDI. Voluntary activation of the PFM elicits the isolated-pelvic coordination pattern which does not have co-activation of the GMM or FDI. Voluntary activation of the GMM activates the gluteal-pelvic coordination pattern which co-activates the GMM and PFM. The inset shows one example voluntary GMM activation, demonstrating that coordinated PFM activity occurred in advance of GMM activity. (B) Example of EMG amplitude and latency in a single participant as voluntary GMM activation is increased. Since all participants performed repeated and scaled muscle contractions, each participant provided EMG data that were used in the group analyses. (C) left. Population average (solid line, ± SEM) PFM activation – separately during voluntary GMM activation and voluntary FDI activation – from a linear mixed model with participant-level random intercepts and slopes. Increases in PFM activation associated with increases in voluntary GMM activation were significantly larger than those associated with increases in voluntary FDI activation (p < 0.005). (C) right. Population average (solid line, ± SEM) PFM activation latency relative to the onset of the GMM – during voluntary GMM activation – from a linear mixed model with participant-level random intercepts and slopes. PFM activation led GMM activation across the range of voluntary GMM activation, with slightly shorter latencies for higher levels of voluntary GMM activation (p = 0.03).

Figure 2

Functional magnetic resonance imaging (fMRI) evidence of non-localized control of human pelvic floor muscles (PFM). (A-1) fMRI data were collected while a first group of participants (Group A) voluntarily activated the first dorsal interosseous muscle (FDI) and the gluteus maximus muscle (GMM) in blocks at a comfortable level. (A-2,3) Contrast of voluntary GMM activation greater than voluntary FDI activation produced significant brain activation in distinct brain areas (Z > 2.3, p < 0.05, cluster corrected), including three distinct ROIs in the motor cortex: supplementary motor cortex (SMA) and bilateral primary motor cortex (M1-L and M1-R). (B-1) fMRI data were collected while a different set of participants (Group B) voluntarily activated the GMM and PFM in blocks to three distinct levels: low (L), medium (M), and high (H). (B-2) The motor cortical ROIs (SMA, M1-L, M1-R) were projected into the native space to identify voxels in these ROI that were active (Z > 2.3, p < 0.05, cluster corrected) in the contrast of muscle activity compared to rest, regardless of contraction level. For each participant and ROI, BOLD signals extracted from the identified voxels were averaged. (B-3) Activation significantly increased in all ROIs during increased voluntary GMM activation (gluteal-pelvic coordination pattern) and during increased voluntary PFM activation (isolated-pelvic coordination pattern). Specifically, activity significantly increased in SMA during increased voluntary GMM (p < 0.0001), in M1-L during increased voluntary GMM (p < 0.0001), in M1-R during increased voluntary GMM (p < 0.0001), in SMA during increased voluntary PFM (p < 0.0001), in M1-L during increased voluntary PFM (p < 0.0001), and in M1-R during increased voluntary PFM (p < 0.0001).

Figure 3

Transcranial magnetic stimulation (TMS) evidence of distributed motor cortical representation of human pelvic floor muscles (PFM). (A) Individual data from 7 participants who had both functional magnetic resonance (fMRI) and TMS testing. First row of images shows fMRI activation maps for the contrast PFM greater than rest (Z > 2.3, p < 0.05, corrected). Notice the distributed activation extending posterior and lateral to medial supplementary motor area (SMA), consistent with primary motor cortex (M1). Second row of images shows TMS maps of motor evoked potentials (MEPs) generated in the PFM during stimulation at rest. Notice again the MEPs did not appear confined to the medial wall and SMA. Shapes show locations chosen by the experimenter during the TMS session for further analysis: significant MEP and overlying SMA (circles), significant MEP and overlying left M1 (triangles), and a non-significant MEP reference site (hexagon). Subsequent rows show MEPs (averaged across 3 stimuli) evoked in the PFM from these locations in these participants. Notice that MEPs appeared in the PFM 21 milliseconds (ms) after the TMS pulse, consistent with direct (though not necessarily monosynaptic) connections, and that MEPs evoked from M1 were not generally smaller – and in some cases larger – than MEPs evoked from SMA. (B) Statistical analysis of MEP peak-to-peak size (normalized to SMA) revealed that PFM MEPs induced by stimulating M1 were significantly larger than PFM MEPs induced by stimulating the reference (p < 0.001), but were not significantly different from PFM MEPs induced by stimulating SMA (p > 0.05). (C) The TMS map for the PFM (averaged across participants) broadly mirrored the fMRI group activation map just under the scalp (z = 66 mm, MNI coordinates) for the contrast of PFM greater than first dorsal interosseous (FDI). Example coordinates chosen according to the Jülich anatomical atlas consistent with SMA (circle), left M1 (triangle) and right M1 (square) are shown for reference.

Figure 4

Transcranial magnetic stimulation (TMS) evidence of increasing neural activation in human supplementary motor area (SMA) and primary motor cortex (M1) during increased voluntary drive. (A) Pelvic floor muscles (PFM) hotspots in SMA and M1-L were separately stimulated to test whether MEPs can be elicited in the PFM (top row) and the gluteus maximus muscle (GMM, bottom row) at rest (first column), during PFM voluntary activation (second column), and during GMM voluntary activation (third column). Plotted MEPs were limited to contraction levels ≤ 25% MVC. PFM MEPs were observed at rest and during PFM and GMM voluntary activations. However, no GMM MEPs were observed at rest or during PFM voluntary activation. GMM MEPs were only observed during GMM voluntary activation. (B) Population average (solid line, ± SEM) MEPs peak-to-peak values (amplitudes) from a linear mixed model with participant-level random intercepts and slopes. In both motor regions, increased PFM activation resulted in significantly increased amplitudes of PFM MEPs (PFM-SMA: p = 0.0001, PFM-M1-L: p = 0.0017). However, increasing PFM activation did not result in increasing either GMM activation level or GMM MEPs amplitudes. In both motor regions, increasing GMM activation resulted in modulating PFM activation level, as well as, in modulating amplitudes of PFM and GMM MEPs (PFM-SMA: p = 0.0076, PFM-M1-L: p = 0.0011, GMM-SMA: p = 0.0052, GMM-M1-L: p = 0.009). (C) Distribution of the difference between onset latency of PFM MEPs and onset latency of corresponding GMM MEPs during GMM voluntary activation. Using linear mixed model with participant-level random intercepts and slopes showed that increasing GMM activation was not associated with changes in the difference (latency of PFM MEPs - latency of GMM MEPs, SMA: p = 0.9539, M1-L: p = 0.3954) that remained about +3 ms on average and was significantly greater than zero (SMA: p = 0.0121, M1-L: p = 0.0114).

Figure 5

Schematic representation of proposed cortical representation of the human pelvic floor. We proposed that, similar to other muscles such as the gluteus maximus muscle (GMM), pelvic floor muscles (PFM) have a distributed representation in the supplementary motor area (SMA) and the primary motor cortex (M1). We further propose that the PFM representation has a lower activation threshold than synergistic muscles such as the GMM. This activation threshold difference may explain why the PFM is activated in advance of the synergists during cortically-driven activation. Mn = Motor neuron.