Assessing the digit organisation of focal hand dystonia using 7T functional MRI

Abstract

Introduction:

Focal hand dystonia (FHD) is typically treated with Botulinum Toxin (BTX/BoNT-A/Botox/Onabotulinum Toxin Type A) injected into digit and wrist muscles to provide temporary symptomatic relief from muscle spasm. We use 7T functional MRI (fMRI) to study 1) differences in somatosensory and motor cortical digit maps between FHD individuals and age and sex-matched healthy volunteers (HV), 2) changes in cortical digit maps of FHD individuals after BTX treatment.

Methods:

Four FHD patients underwent behavioural measures and 7T fMRI at two timepoints: ∼30 days after BTX treatment at peak efficacy (Post-BTX), immediately prior to their next injection ∼100 days after treatment when therapeutic effects had declined (Pre-BTX). HVs were scanned at a single time point. Behavioural measures included temporal discrimination (TDT), amplitude thresholding (AMP) and spatial acuity grating orientation tasks. A somatosensory travelling wave (TW) digit-mapping fMRI task was performed on dominant/affected and non-dominant/un-affected hands, and a motor TW digit-mapping task on the dominant/affected hand. Fourier analysis derived digit maps which were compared to a probabilistic digit atlas, and population receptive field (pRF) mapping performed. Structural and resting state fMRI data were acquired.

Results:

Spatial acuity was lowered in the dominant/affected hand in FHD compared to HVs, while TDT and AMP measures showed no significant differences. Somatosensory and motor TW cortical digit maps in FHD and HV showed clear digit ordering. Somatosensory and motor cortical digit maps were more disordered in the dominant/affected hand than the non-dominant hand for FHD Post and Pre-BTX compared to HVs (Dice coefficient/Figure of Merit). Compared to the atlas, HV maps were in-line for both hands, while in FHD central tendency (CT) was lower in the dominant/affected hand than non-dominant hand. pRF sizes were larger in the FHD Pre-BTX group compared to HV and Post-BTX for the dominant hand.

Conclusion:

Behaviourally, FHD patients had lower spatial acuity than HVs. FHD digit maps had typical D1-D5 ordering, but Dice coefficients identified a disordered dominant hand in FHD patients compared to HVs at both timepoints. The FHD Pre-BTX group had larger pRF sizes than Post-BTX and HVs, suggesting BTX treatment coincides with more localised touch perception.

Introduction

Focal Hand Dystonia is a sensorimotor disorder presenting with involuntary movement and tremors in the hand [1]. Early brain imaging studies of focal hand dystonia (FHD) have suggested disordered somatosensory function demonstrated by reduced distances between the cortical representation of digits and increased overlapping activity compared to healthy controls [2, 3]. However, because the cortical digit maps of somatosensation and motor function representing the hand are at the millimetre scale, these findings of disorder are now being questioned as higher spatial resolution fMRI become available, motivating this work.

Functional MRI (fMRI) studies probing the cortical digit representation use block [4, 5], phase-encoding [6] and event-related [7] designs to deliver sensory stimulation to the fingers, mostly in healthy individuals. In the past decade, there has been a push to investigate digit maps at higher magnetic field strength, due to the increased BOLD contrast and spatial resolution [8, 9]. Somatotopic digit mapping in human primary somatosensory cortex using a travelling-wave (TW) phase-encoding fMRI paradigm has been shown to be highly reproducible in healthy participants at 7T [10] but show inter-participant variability. To address this, we developed a probabilistic somatosensory atlas [11] to define digit dominance and inter-participant variability in healthy participants. This atlas provides a standardised template for comparison with patient groups, such as those people with Focal Hand Dystonia (FHD).

Prior studies in dystonia have principally used motor tasks. A study in dystonic musicians [12] showed greater activity in ipsilateral premotor area and reduced activity in the cerebellum during a right-handed tapping task, suggesting irregular neural activity and compensatory mechanisms compared to healthy controls. In contrast, a study of a cued movement task in FHD patients [13] showed a decrease in ipsilateral activity and an increase in contralateral activity associated with the affected hand. A further study [14] showed that motor performance to a button press task was similar between controls and FHD patients, while the motor preparation section of the task showed reduced activation in premotor and postcentral gyrus for FHD patients, implying impaired early stages of motor control. A 3T fMRI study using passive extension and active finger presses of individual fingers in musician’s dystonia (n = 9) [15] discerned individual digits and clear somatotopy but showed no difference in the spatial geometry (expansion or spatial shift of digit representation) for musicians with and without dystonia. This suggestion of the intact cortical digit maps in musician’s dystonia contests the idea that task-specific dystonia alters cortical maps and lends weight to an alternative hypothesis that task-specific dystonia is due to a higher-order disruption of skill encoding. A recent high spatial resolution (0.75 × 0.75 × 0.89 mm) laminar 7T MRI Vascular Space Occupancy (VASO) study showed in a small group of individuals with FHD (n = 4) a breakdown of digit somatotopy and increased activation in superficial cortico-cortical input layers during finger tapping compared to unaffected hemispheres, suggested to be due to reduced inhibition [16].

Together, these studies of FHD suggest widespread differences in sensorimotor areas and subcortical structures. However, these prior studies have generally been performed at a field strength of 3T or less and have used spatial analyses (e.g., centre of gravity of each digit based on vertices and searchlight analysis) in volumetric space. There is therefore a need for high spatial resolution high field fMRI studies using surface-based methods to provide more accurate and more sensitive results [17] and better digit delineation across gyral and sulcal boundaries. More recently, fMRI studies have measured how receptive fields of neurons in the somatosensory cortex are distributed, for example, by using a 2D grid of stimulators to map within- and between-digit characteristics. Computational models have been used to investigate more fine-grained characteristics of the hand, such as population receptive field (pRF) size, in healthy participants [18–21].

Resting state fMRI (rs-fMRI) has also shown alterations in functional connectivity (FC) in dystonia. In Writer’s Cramp (WC) [22] reduced functional connectivity compared with controls was shown when placing seeds in left lateral premotor cortex, left thalamus, left/right pallidum to study the projection to the symptomatic left primary sensorimotor cortex. There were also reductions in FC from primary motor cortex premotor, frontal and somatosensory areas. Reductions in FC were also shown in cervical dystonia, in regions related to the M1-SMA motor network [23]. Another study showed abnormal recruitment of parietal and premotor cortices, as well as distinct patterns of sensorimotor integrations between musician’s and non-musician’s dystonia [24]. Unique patterns for dystonic tremor, based on a reduction in FC were found in higher-level cortical, basal ganglia, and cerebellar regions [25]. Further, a study of cervical dystonia and blepharospasm [26] suggested that focal dystonia is a disorder of specific networks and not just a result of basal ganglia alterations.

Behaviourally, there is evidence of raised temporal thresholds in patients as determined with temporal discrimination tasks and spatial discrimination thresholds using a grating orientation task (GOT) [27–29].

A symptomatic treatment of FHD involves injections of Botulinum Toxin (BTX), into the muscles controlling digit and/or wrist movement. This can induce relief from involuntary muscle spasm associated with FHD. A prior fMRI study in patients with cervical dystonia affecting the neck showed an increase in activation in sensorimotor cortex, 4 weeks after BoNT injection [30].

Here we used high spatial resolution 7T fMRI and surface-based analyses to study cortical somatotopy digit maps and the pRF digit size and perform rs-fMRI in a group of FHD patients and healthy controls (HV). Patients were scanned at two timepoints in the BTX treatment cycle of, ∼30 days after localised BTX injection at peak efficacy (Post-BTX), and immediately prior to their next injection ∼100 days after treatment when therapeutic effects have declined (Pre-BTX). At each timepoint, behavioural tasks of temporal discrimination, amplitude thresholding and spatial acuity were collected.

For these measures we predict FHD responses at 4 weeks following BTX treatment are more like those of healthy controls compared those at 3 months following BTX treatment.

Materials and methods

The study was conducted with the approval of the University of Nottingham Faculty of Medicine and Health Sciences Ethics Committee for healthy volunteers and of the Health Research Authority Research Ethics Committee (REC 17/EM/0368) for FHD patients. FHD patients were recruited from the Botulinum Toxin Clinic at Queens Medical Centre. Eligibility criteria included patients 18 and 75 years who were clinically assessed to have FHD and presented with unilateral symptoms on their dominant hand. Exclusion criteria included participants with a history of neurological illness other than FHD and MRI contraindications. All participants gave written informed consent and completed a MR safety screening form and a handedness form at each visit.

Eight patients (2 visits) with Focal Hand Dystonia (FHD) (age mean ± stdev: 54 ± 14 years, 4 male) were recruited to be scanned at 7T at both 4 weeks (20–27 days) (Post-BTX) and 3 months (105–119 days) after (Pre-BTX) Onabotulinumtoxin type-A (Botox, BTX) treatment [31] and age-matched healthy volunteers (HVs) (age mean ± stdev: 42 ± 7 years were scanned. One patient had a recent diagnosis of writer’s cramp, and the study was performed following their second set of BTX treatment. All other participants had longstanding writer’s cramp with dystonic posturing but no tremor and were having ∼3 monthly Botox treatment. All patients reported good therapeutic response, although Participant P03 reported mild weakness of thumb flexion following treatment.

At each visit, participants completed 1 h of behavioural data collection followed by 7T functional and structural MRI scanning lasting 1 h. In all FHD participants, the affected hand was also the dominant hand.

Behavioural data acquisition

Behavioural data collection included a temporal discrimination task (“TDT”) and an amplitude threshold (“AMP”) task coded in MATLAB (The Mathworks, Natick, Massachusetts), and spatial acuity grating orientation task (“GOT”). These tasks were performed using identical piezoelectric stimulators to those used in the fMRI study, while the GOT used Perspex domes.

The “TDT” was performed on the tips of digits D2 and D3 of left and right hands using both piezo-electric stimulators and a Brain Gauge device [32]. For the “AMP” task, participants completed a two-interval forced choice amplitude detection task at 30 Hz (match fMRI paradigm) and 200 Hz on the left and right index finger (D2). The “GOT” was used to assess spatial acuity [33] of the tip of D2, due to time constraints only one hand was assessed: the affected hand in FHD patients and the dominant hand in controls (see Supplementary Material for further details).

In addition, 10 healthy participants (age mean ± stdev: 36 ± 11 years) were recruited to assess the repeatability of the behavioural tasks. The participants underwent the same procedure as those in the main study (TDT using piezo, AMP and GOT). The tasks were each performed twice, separated by a two-week interval. This allowed both the within- and between-participant coefficient of variation (CV) of measures to be computed, and the comparison of the CV to Mikkelsen et al. [34] who assessed reproducibility of vibrotactile detection and discrimination.

MRI data acquisition

Data were collected on a 7 T Philips Achieva scanner using a 32-channel receive coil (Nova Medical). Travelling wave (TW) fMRI data was collected to assess somatotopy and motortopy. Somatosensory stimuli were delivered to the fingertips (D1/D2/D3/D4/D5) using MR-compatible piezo-electric stimulators driven at 30 Hz vibrotactile frequency (Dancer Design, UK; http://www.dancerdesign.co.uk/). Motortopy was performed in the dominant hand. Additionally, resting state fMRI and structural MRI data were collected (see Supplementary Material for the full MR acquisition protocol).

MRI analysis

Following quality control (tSNR>40, movement <1 mm), only four participants were included going forward. fMRI data were distortion-corrected (FSL TOPUP), motion-corrected and aligned to the FLASH image. Data were high-pass filtered with a 0.01 Hz cutoff (100 s) to remove low-frequency trends. fMRI data were then analysed using custom software within mrTools [35] implemented in MATLAB. We performed several analyses to study the spatial representation of the digits and compare FHD patients to healthy controls, as described in Sections Somatotopy and motortopy phase analysis of TW data, Comparison of data to the somatosensory probabilistic atlas, Population receptive field (pRF) mapping, Resting state connectivity, Data and code availability statement. We also confirmed our digit representations from phase analysis of the TW data against a General Linear Model (GLM) analysis to confirm correspondence between analysis methods (Full details are provided in Supplementary Material).

Somatotopy and motortopy phase analysis of TW data

Somatotopy and motortopy data were analyzed using standard, Fourier based methods (in mrTools), producing coherence and phase maps. To account for shifts in the fMRI response due to the haemodynamic lag, the forward and reverse scans were averaged, as described in [7]. Coherence, phase, and amplitude of the best-fitting sinusoid were then computed as described in [36]. The phase map was displayed on the flattened cortical patch with a threshold of c > 0.3 (coherence); this corresponds to an uncorrected p value of p < 0.0004. First the digit correspondence between the somatosensory task data and the probabilistic Atlas were compared using Dice coefficients in fsaverage space for healthy volunteers and FHD patients for each visit. Then digit correspondence of the somatosensory and motor data was assessed using Dice coefficients between the HVs and FHD patients for each visit (Post-BTX and Pre-BTX), and between FHD Post-BTX and Pre-BTX. For each group, we also assessed Dice coefficients related to the extent of correspondence of motor and somatosensory maps in the Dominant hand. For the comparison between HV and FHD patients the Figure of Merit (FoM) was calculated from the Dice scores to summarise the response of each digit; this metric penalises degeneracy and poor representation [37].which defines the Dice (r) of the participant’s digit ROI (m) to its paired ROI (X) subtracting the mean of the Dice from the other digit ROIs. A FoM score of 1 represents perfect overlap with the target digit, whilst a lower score suggests the digit overlaps with the non-target digit or the representation is degenerate (i.e., overlaps with many digits).

Comparison of data to the somatosensory probabilistic atlas

ROIs of each digit from the TW phase analysis were normalised to standard-space (fsaverage) and compared to the somatosensory probabilistic atlas [11]. Surface projection and normalisation followed the methods described in O’Neill et al., i.e., using Freesurfer’s mri_vol2surf to project coherence maps, phase maps, and digit ROIs into the participant’s surface space, followed by a surface-to-surface transformation to MNI-305 space with mri_surf2surf. Since surface-based registration has previously been shown to outperform volume comparisons [11], only surface-based transformations were performed in this work. Full probabilistic maps (FPMs) were defined for each group (HC, and FHD for both Post-BTX and Pre-BTX), in which probabilities at each vertex were defined as the number of participants for which a particular digit was assigned to this voxel/vertex, divided by the total number of participants.

For the somatosensory data, central tendency (CT) was measured between the FHD patients’ digits and the atlas digits. CT defines how centrally located a digit is relative to a probabilistic atlas digit (1 = perfect overlap, >1 = resides centrally, <1 = resides peripherally) [38]. CT quantifies how much a single participant’s digit ROI overlaps with high probability areas of the atlas. For example, a small focal ROI that is within the atlas ROI would give a high central tendency. A larger ROI that encompasses the atlas ROI would also give a large central tendency if the two centres were aligned.

Population receptive field (pRF) mapping

This analysis followed the methods described in [21] for somatosensory and motortopy tasks, which was adapted from a pRF method developed for visual cortex [39]. Briefly, a 1-dimensional Gaussian model was combined with the spatio-temporal description of the stimulus, producing a predicted “cortical” response. To account for haemodynamic effects, this was then convolved with a HRF and then fit to the measured timeseries using non-linear least squares optimisation. This method produces cortical maps of “between digit” (equivalent to the TW Fourier analysis, albeit parameterized), pRF “size” (defined from the standard deviation of the 1D Gaussian best fit for each voxel) and adjusted r² (the goodness of fit of the pRF model to the fMRI timeseries, pRF outputs were threshold by adjusted r² > 0 (since adjusted r² can be negative). ROIs were calculated by intersecting the binned digit phase value (D1-D5) with the FreeSurfer Brodmann areas (3a,3b,1,2), creating 20 ROIs for each participant. pRF size was extracted from each ROI and normalised (converted to z-scores) to account for inter-participant variability in absolute pRF size. All statistical comparisons were Bonferroni corrected for multiple comparisons at α = 0.05.

Resting state connectivity

Data were analysed using the CONN toolbox in MATLAB (RRID:SCR_009550) [40] (Whitfield-Gabrieli and Nieto-Castanon 2012). First, the default preprocessing pipeline was used which included realignment, outlier detection, co-registration, normalisation to the MNI template, and 5 mm spatial smoothing. Denoising was performed using aCompCor [41], scrubbing, and temporal band pass filtering [0.008 0.09Hz]. Seed-based functional connectivity (FC) measures to cortical brain regions were then assessed by placing seed ROIs in both the left (L) and right (R) post-central gyrus taken from the Harvard-Oxford atlas (“atlas.PostCG r” and “atlas.PostCG l”). FC was represented by Fisher-transformed bivariate correlation coefficients from a weighted GLM, defined separately for each pair of seed and target areas, modelling the association between their BOLD signal timeseries. Cluster-level inferences were based on parametric statistics from Gaussian Random Field Theory; results were threshold using a cluster-forming p < 0.001 voxel-level threshold, and familywise corrected p-FDR <0.05 cluster-size threshold. Group independent component analysis (groupICA) was performed. FC was compared for 1) HC vs. FHD (for both Post-BTX and Pre-BTX), and 2) FHD between visits (Post-BTX vs. Pre-BTX).

Data and code availability statement

Code to generate results and figures, and the accompanying derived data is freely available on the Mendeley data repository: https://data.mendeley.com/datasets/tvt7mg5bf3/1.

Results

Behavioural measures

Spatial acuity using the “GOT” task was significantly reduced in FHD for both Post-BTX and Pre-BTX groups compared with healthy controls (Mean HV (std): 1.87 (0.19), Mean Post-BTX (std): 2.96 (0.93), Mean Pre-BTX (std): 3 (0.88); one-way ANOVA p < 0.05, Bonferroni corrected), Figure 1A.

FIGURE 1

34]; (C) Temporal discrimination task (“TDT”) using piezoelectric stimulators and Brain Gauge (BG), and (D) Amplitude discrimination (“AMP”) at 31 and 200 Hz.

In the healthy control repeatability group, within-participant CV for behavioural measures were: “GOT”: 6.2% ± 1.7%; “TDT”: 40% ± 12% (R) 46% ± 11% (L) (Mikkelson: 46%); “AMP”: 21 % ± 6% (R) 28% ± 8% (L) (Mikkelson: 31%), as shown in Figure 1B and Supplementary Table S1 which provides between-participant CVs. These CVs are in-line with those reported by Mikkelson [34].

There were no significant differences in the “TDT” tasks or “AMP” task between the FHD group and healthy controls, or for FHD patients between Post-BTX and Pre-BTX visits (Figures 1C,D).

Comparing travelling wave digit maps between HV and FHD patients

Qualitatively, FHD travelling wave somatotopy and motor phase analysis digit maps showed similar digit-specific activity patterns to HVs, with FHD patterns similar between Post-BTX and Pre-BTX visits, Figure 2.

FIGURE 2

Quantitatively, Dice coefficient matrices of somatosensory phase analysis digit correspondence assessed in fsaverage space between groups are shown in Figure 3. When comparing matrices between HV and FHD groups for the somatosensory task (Figure 3Ai) and motor task (Figure 3Aii), lower Dice coefficients were seen in the dominant affected hand compared to the non-dominant hand for HV versus Post-BTX and HV versus Pre-BTX (P < 0.02). This is shown in Figure 3Aiv) by the lower FoM of the Dice scores in the Dominant affected hand. However, no significant difference in the Dice matrices was detected between Post-BTX and Pre-BTX conditions. Figure 3Aiii is a graphic that explains how the Dice matrix affects the FoM score.

FIGURE 3

In response to the somatosensory task the dominant hand had larger digit ROIs than non-dominant digit ROIs for FHD and HV groups (P < 0.03), with D5 being significantly smaller than all other digit (D1-D4) ROIs (P = 0.004) (see Supplementary Figure S1).

Somatosensory and motortopy task dominant hand digit maps showed strong spatial correspondence, with clear diagonal high Dice coefficients in each HV, FHD Post-BTX, FHD Pre-BTX group (Supplementary Figure S2). For all participants, there was good agreement in digit correspondence of the TW GLM “winner-take-all” maps with the standard phase analysis maps shown by the high diagonal Dice matrix coefficients (Supplementary Figure S3).

Comparing digit maps to the probabilistic atlas

The Dice of the TW phase analysis digit maps in response to the somatosensory task with the Atlas digit ROIs showed a trend for a stronger diagonal in the HV compared to FHD group, particularly for the Dominant (affected) hand, Figure 4A (mean Dice diagonal: HV L/R: 0.634 ± 0.009/0.62 ± 0.02; Post-BTX L/R: 0.61 ± 0.02/0.60 ± 0.01; Pre-BTX L/R: 0.58 ± 0.02/0.58 ± 0.01). This is also reflected in the central tendency (CT) matrices (Figure 4B) (mean CT diagonal: HV L/R: 1.1 ± 0.1/1.2 ± 0.3; Post-BTX L/R: 1.3 ± 0.2/1.0 ± 0.2; Pre-BTX L/R: 1.3 ± 0.2/0.9 ± 0.2) which also considers how centrally located a digit is. A within group paired t-test between Dominant and Non-dominant hand CT scores, showed that for both FHD visits (Post-BTX and Pre-BTX) the Dominant hand had significantly smaller CT values than the non-Dominant hand (p < 0.05, Bonferroni corrected).

FIGURE 4

Population receptive field (pRF) mapping

Mean (absolute) pRF size (across hemispheres and digits) across groups was HV (1.7 ± 1.8), HV Atlas (1.4 ± 1.2), FHD Post-BTX (2.0 ± 1.9), FHD Pre-BTX (1.7 ± 1.9), these results are in-line with our previous study in HVs (1.9 ± 1.6 2D Gaussian model, mean ± stdev) [21].

Comparing group mean pRF sizes for the somatosensory task, the FHD Pre-BTX group had larger pRF sizes in the Dominant (affected) hand than the FHD Post-BTX, HV, and Atlas. There were no differences in pRF size between groups in the Non-dominant hand for the somatosensory task, or in the motor pRF data for the Dominant hand (Figure 5). There were no differences in pRF sizes from the motortopy data in the Dominant hand (the only hand on which the motor task was performed). There was no overall significant difference in pRF size between hands across groups. Motor pRF size was larger than somatosensory pRF size for both HV and FHD groups (Supplementary Figure S4).

FIGURE 5

Table 1 and Figure 6 show the pRF size ordering across digits in each group for non-dominant and dominant hands for the somatosensory task and dominant hand for the motor task. While HV and FHD Post-BTX showed greater pRF sizes for D2,3,4,5 than D1 this was not seen for FHD Pre-BTX in the dominant hand. In the non-dominant hand, all groups had similar pRF size trends, e.g., D2,3,4 > D1. Table 1 also shows pRF results separated by Brodmann areas, with BA2 larger than BA3a and 3b in all groups in the dominant hand, however, in the non-dominant hand, only Post-BTX and Atlas groups show BA2 > BA3b.

FIGURE 6

Resting state fMRI

There was a significant (p < 0.001, uncorrected) difference in seed-based connectivity between the seed in left postcentral gyrus (contralateral to dominant hand) and projection to right postcentral gyrus, with lower connectivity in FHD, Figure 7. There was no difference in FC in FHD between Post-BTX and Pre-BTX visits.

FIGURE 7

Discussion

This work assessed behavioural and functional measures of digit localisation and population receptive field (pRF) size using high spatial resolution fMRI in FHD patients compared to healthy volunteers, and the effect of treatment with BTX.

Behavioural temporal and spatial thresholds in FHD

The larger spatial acuity in FHD patients using the “GOT” task agrees with previous literature [28] and aligns with broader evidence of sensory processing deficits in FHD (for a review, see [42]). While other studies have reported prolonged TDTs in FHD patients [29, 43, 44], this has been shown to be dependent on the experimental protocol [45]. Here, the mean TDT metrics agreed with previous literature [34], but did not show any significant groups differences. This may be due to sample size and the large variability in the “TDT” measure with the “GOT” task having higher robustness in FHD [45].

Dominant (affected) and non-dominant (unaffected) hand digit maps in FHD

Our findings of preserved digit maps for travelling wave or GLM analyses contest early work showing widespread digit disruption in FHD [2, 46], but supports more recent work showing intact digit representation within primary sensorimotor cortex of individuals with musician’s dystonia [15]. However, the lower Dice coefficient in the FHD affected hand compared to HVs, points to a degree of disorganisation or increased variability in the digit representations in the affected hand likely due to abnormal sensorimotor integration [28]. It should be noted that early fMRI studies [2] were performed at coarser spatial resolution and assessed digit maps in un-flattened cortical space based on Euclidean distances between digits, so differences may have arisen due to folding structure or limitations of spatial resolution or altered amplitude of responses in dystonia patients.

Digit 5 (D5) was significantly smaller than all other ROIs, agreeing with literature on this digit having the least representation in terms of size (for an overview of cortical magnification of digits, see [47]).

We note large overlap between somatosensory and motor digit maps, evidenced by Dice coefficients, which reflects the anatomical and functional relationship of primary S1 and M1, which form an integrated sensorimotor representation of the digits [48].

Comparison of central tendency scores of digit maps to the probabilistic atlas

In FHD patients, CT results showed the dominant hand has larger less centrally matched digit representations, whilst the non-dominant hand has smaller more centrally represented digits. It is noted that the differences between groups may be due to atypical HVs, however, when comparing groups to the Atlas, the HVs perform equal to or better than FHD patients (Figure 4B).

Population receptive field sizes in FHD

The dominant hand showed greater cortical disruption at Pre-BTX than Post-BTX, with larger pRF sizes and lower DICE scores, suggesting BTX treatment transiently affected the cortical representation of the digits (see Section Effect of Botulinum Toxin).

The thumb had the smallest pRF size (D2, D3, D4, D5 > D1). This ordering is consistent with electrophysiological and primate literature, whereby D1 has the highest receptor density, smallest receptive fields compared to the other digit tips and the largest cortical magnification [49, 50].

Previous fMRI pRF studies have also shown D5 to be larger than the other digits. Interestingly, we showed in some cases that the pRF size of D3 and D4 was greater than D5, where other studies have found D5 to have the largest pRF density size. This may be due to fewer good fits of the pRF fitting algorithm at D5 [47] as less volume is dedicated to D5 in the somatosensory hand map [51].

Our finding of no significant difference between dominant and non-dominant hands in digit representations and pRF size in our healthy participants agrees with previous work which showed no significant differences in volume, activation strength, and Euclidean distance between hands [52], or tactile acuity [53].

Consistent with previous literature that showed larger pRF sizes in precentral gyrus [18], motor pRF sizes, measured across both post-central and pre-central gyrus, were larger than somatosensory pRF size.

Resting state functional connectivity in FHD

Reduced functional connectivity (FC) in FHD is consistent with previous reports of lower sensorimotor resting-state connectivity in dystonia [22, 54], however, those studies included larger cohorts (n = 16, 15, respectively). The absence of FC differences between Post-BTX and Pre-BTX visits likely reflects limited statistical power given the small sample, rather than a true absence of treatment effect.

Effect of botulinum toxin

Patients reported good therapeutic response, with BTX treatment being associated with changes in pRF size. Specifically, pRF size was smaller at Post-BTX (∼30 days) than Pre-BTX (∼100 days), and comparable to HVs at Post-BTX. In contrast, spatial acuity “GOT” measures while raised in patients did not differ with BTX treatment, suggesting the tuning of cortical representations using pRF size captures a more sensitive aspect of somatosensory organisation than behavioural measures. Dice coefficients and CT metrics did not differ between Post-BTX and Pre-BTX, suggesting that more large-scale digit ordering is preserved.

The increase in pRF size at Pre-BTX, implies a reversible, treatment-associated modulation of somatosensory cortical organisation consistent with findings in cervical dystonia, where increased sensorimotor cortical activation was observed 4 weeks post-BTX injection, though the mechanisms may differ given the peripheral versus central site of dystonia [30]. Clinically, BTX acts peripherally by reducing abnormal muscle activity, yet pRF changes suggest an additional central mechanism; consistent with sensory processing deficits in FHD [28, 47].

Association between behavioural and fMRI measures

The combination of larger spatial acuity (“GOT”), lower Dice coefficients and CT scores, and larger pRF sizes in the dominant/affected hand suggest these measures reflect a subtle disruption in spatial cortical organization in FHD. Assessing the correlation between the “GOT” metric and fMRI measures showed a positive association with pRF size (r = 0.498, p = 0.099), but no association with mean CT diagonal (r = −0.317, p = 0.316) or mean Dice diagonal (r = −0.096, p = 0.766).

Limitations

The main limitation of this study is the small sample size which reduces statistical power. Within-participant variability across BTX treatment visits (e.g., injection site variation, individual treatment response) and between-participant heterogeneity in FHD symptoms (e.g., affected digit, duration, severity) complicate group-level interpretation.

Given the small sample size, we acknowledge that multiple comparisons made across digits, Brodmann areas, dominance, and groups increases the risk of false positives despite all group-level comparisons being multiple comparisons corrected.

The “GOT” (spatial acuity) was only performed on the dominant/affected hand due to time constraints, negating comparisons between hands (within the same participant). Similarly, the motor task was only acquired on the dominant hand.

Future work should address these limitations with a larger cohort and perform motor mapping of both hands for comparison of dominant and non-dominant hands.

Conclusion

Spatial acuity was raised in FHD patients compared to healthy controls. The FHD group had some alterations in their Dominant affected hand with both lower Dice scores and larger pRF sizes between Post-BTX and Pre-BTX. These results provide some evidence to support a neuronal change in the cortical representation of the affected hand in focal hand dystonia and that BTX treatment in FHD patients can alter pRF size in the dominant affected hand.

References

• 1.

  The Dystonia Society. Focal Hand Dystonia (2014). Available online at: http://www.dystonia.org.uk/index.php/professional-research/types-of-dystonia/focal-hand-dystonia (Accessed July 20, 2008).

  • Google Scholar
  • View reference in article

• 2.

  ButterworthSFrancisSKellyEMcGloneFBowtellRSawleGV. Abnormal cortical sensory activation in dystonia: an fMRI study. Move Disord (2003) 18(6):673–82. 10.1002/mds.10416

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 3.

  NelsonAJBlakeDTChenR. Digit-specific aberrations in the primary somatosensory cortex in writer’s cramp. Ann Neurol (2009) 66(2):146–54. 10.1002/ana.21626

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 4.

  MartuzziRvan der ZwaagWFarthouatJGruetterRBlankeO. Human finger somatotopy in areas 3b, 1, and 2: a 7T fMRI study using a natural stimulus. Hum Brain Mapp (2014) 35(1):213–26. 10.1002/hbm.22172

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 5.

  SchweizerRVoitDFrahmJ. Finger representations in human primary somatosensory cortex as revealed by high-resolution functional MRI of tactile stimulation. Neuroimage (2008) 42(1):28–35. 10.1016/j.neuroimage.2008.04.184

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 6.

  Sanchez-PanchueloRMFrancisSBowtellRSchluppeckD. Mapping human somatosensory cortex in individual subjects with 7T functional MRI. J Neurophysiol (2010) 103(5):2544–56. 10.1152/jn.01017.2009

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 7.

  BesleJSanchez-PanchueloRMBowtellRFrancisSSchluppeckD. Single-subject fMRI mapping at 7 T of the representation of fingertips in S1: a comparison of event-related and phase-encoding designs. J Neurophysiol (2013) 109(9):2293–305. 10.1152/jn.00499.2012

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 8.

  SchluppeckDSanchez-PanchueloRMFrancisST. Exploring structure and function of sensory cortex with 7T MRI. Neuroimage. (2016) 164:1–8. 10.1016/j.neuroimage.2017.01.081

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 9.

  van der ZwaagWSchaferAMarquesJPTurnerRTrampelR. Recent applications of UHF-MRI in the study of human brain function and structure: a review. NMR Biomed (2016) 29(9):1274–88. 10.1002/nbm.3275

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10.

  Sanchez-PanchueloRMBesleJMouginOGowlandPBowtellRSchluppeckDet alRegional structural differences across functionally parcellated Brodmann areas of human primary somatosensory cortex. Neuroimage (2014) 93:221–30. 10.1016/j.neuroimage.2013.03.044

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 11.

  O’NeillGCSenguptaAAsgharMBarrattELBesleJSchluppeckDet alA probabilistic atlas of finger dominance in the primary somatosensory cortex. Neuroimage (2020) 217(August 2019):116880. 10.1016/j.neuroimage.2020.116880

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12.

  KadotaHNakajimaYMiyazakiMSekiguchiHKohnoYAmakoMet alAn fMRI study of musicians with focal dystonia during tapping tasks. J Neurol (2010) 257(7):1092–8. 10.1007/s00415-010-5468-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 13.

  KimberleyTJPickettKA. Differential activation in the primary motor cortex during individual digit movement in focal hand dystonia vs. healthy. Restor Neurol Neurosci (2012) 30(3):247–54. 10.3233/RNN-2012-110183

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 14.

  JankowskiJPausSScheefLBewersdorffMSchildHHKlockgetherTet alAbnormal movement preparation in task-specific focal hand dystonia. PLoS One (2013) 8(10):e78234. 10.1371/journal.pone.0078234

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 15.

  SadnickaAWiestlerTButlerKAltenmüllerEEdwardsMJEjazNet alIntact finger representation within primary sensorimotor cortex of musician’s dystonia. Brain (2023) 146(4):1511–22. 10.1093/brain/awac356

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 16.

  HuberLKassavetisPGulbanOFHallettMHorovitzSG. Laminar VASO fMRI in focal hand dystonia patients. Dystonia (2023) 2:10806. 10.3389/dyst.2023.10806

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 17.

  TucholkaAFritschVPolineJBThirionB. An empirical comparison of surface-based and volume-based group studies in neuroimaging. Neuroimage (2012) 63(3):1443–53. 10.1016/j.neuroimage.2012.06.019

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18.

  SchellekensWPetridouNRamseyNF. Detailed somatotopy in primary motor and somatosensory cortex revealed by Gaussian population receptive fields. Neuroimage (2018) 179(May):337–47. 10.1016/j.neuroimage.2018.06.062

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19.

  PuckettAMBollmannSJundayKBarthMCunningtonR. Bayesian population receptive field modeling in human somatosensory cortex. Neuroimage (2020) 208(March 2019):116465. 10.1016/j.neuroimage.2019.116465

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 20.

  WangLZhangZOkadaTLiCChenDFunahashiSet alPopulation receptive field characteristics in the between- and within-digit dimensions of the undominant hand in the primary somatosensory cortex. Cereb Cortex (2021) 31(10):4427–38. 10.1093/cercor/bhab097

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 21.

  AsgharMSanchez-PanchueloRSchluppeckDFrancisS. Two-dimensional population receptive field mapping of human primary somatosensory cortex. Brain Topogr (2023) 36(6):816–34. 10.1007/s10548-023-01000-8

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 22.

  DreselCLiYWilzeckVCastropFZimmerCHaslingerB. Multiple changes of functional connectivity between sensorimotor areas in focal hand dystonia. J Neurol Neurosurg Psychiatry (2014) 85(11):1245–52. 10.1136/jnnp-2013-307127

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 23.

  PanPWeiSOuYJiangWLiWLeiYet alReduced global-brain functional connectivity and its relationship with symptomatic severity in cervical dystonia. Front Neurol (2020) 10:10. 10.3389/fneur.2019.01358

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 24.

  BianchiSFuertingerSHuddlestonHFruchtSJSimonyanK. Functional and structural neural bases of task specificity in isolated focal dystonia. Move Disord (2019) 34(4):555–63. 10.1002/mds.27649

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 25.

  DeSimoneJCArcherDBVaillancourtDEWagle ShuklaA. Network-level connectivity is a critical feature distinguishing dystonic tremor and essential tremor. Brain (2019) 142(6):1644–59. 10.1093/brain/awz085

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 26.

  GiannìCPasquaGFerrazzanoGTommasinSDe BartoloMIPetsasNet alFocal dystonia. Neurology (2022) 98(14):e1499–e1509. 10.1212/WNL.0000000000200022

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 27.

  MolloyFMCarrTDZeunerKEDambrosiaJMHallettM. Abnormalities of spatial discrimination in focal and generalized dystonia. Brain (2003) 126(10):2175–82. 10.1093/brain/awg219

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 28.

  SangerTDTarsyDPascual-LeoneA. Abnormalities of spatial and temporal sensory discrimination in writer’s cramp. Move Disord (2001) 16(1):94–9. 10.1002/1531-8257(200101)16:1<94::AID-MDS1020>3.0.CO;2-O

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 29.

  Bara-JimenezWSheltonPSangerTDHallettM. Sensory discrimination capabilities in patients with focal hand dystonia. Ann Neurol (2000) 47(3):377–80. 10.1002/1531-8249(200003)47:3<377::AID-ANA16>3.0.CO;2-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30.

  NevrlýMHluštíkPHokPOtrubaPTüdösZKaňovskýP. Changes in sensorimotor network activation after botulinum toxin type A injections in patients with cervical dystonia: a functional MRI study. Exp Brain Res (2018) 236(10):2627–37. 10.1007/s00221-018-5322-3

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 31.

  CohenLGHallettMGellerBDHochbergF. Treatment of focal dystonias of the hand with botulinum toxin injections. J Neurol Neurosurg Psychiatry (1989) 52(3):355–63. 10.1136/jnnp.52.3.355

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 32.

  TommerdahlM. Corticalmetrics.com. Brain Gauge (2017). Available online at: https://www.corticalmetrics.com/braingaugeplus (Accessed July 17, 2017).

  • Google Scholar

• 33.

  CraigJC. Grating orientation as a measure of tactile spatial acuity. Somatosens Mot Res (1999) 16(3):197–206. 10.1080/08990229970456

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34.

  MikkelsenMHeJTommerdahlMEddenRAEMostofskySHPutsNAJ. Reproducibility of flutter-range vibrotactile detection and discrimination thresholds. Sci Rep (2020) 10(1):6528. 10.1038/s41598-020-63208-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 35.

  GardnerJLMerriamEPSchluppeckDBesleJHeegerDJ. In: mrTools: Analysis and Visualization Package for Functional Magnetic Resonance Imaging Data (2018). 10.5281/ZENODO.1299483

  • CrossRef
  • Google Scholar

• 36.

  BesleJSanchez-PanchueloRMBowtellRFrancisSSchluppeckD. Event-related fMRI at 7T reveals overlapping cortical representations for adjacent fingertips in S1 of individual subjects. Hum Brain Mapp (2014) 35(5):2027–43. 10.1002/hbm.22310

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 37.

  O’NeillGCTewariePKColcloughGLGascoyneLEHuntBAEMorrisPGet alMeasurement of dynamic task related functional networks using MEG. Neuroimage (2017) 146:667–78. 10.1016/j.neuroimage.2016.08.061

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 38.

  EickhoffSBPausTCaspersSGrosbrasMHEvansACZillesKet alAssignment of functional activations to probabilistic cytoarchitectonic areas revisited. Neuroimage (2007) 36(3):511–21. 10.1016/j.neuroimage.2007.03.060

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 39.

  DumoulinSOWandellBA. Population receptive field estimates in human visual cortex. Neuroimage. (2008) 39(2):647–60. 10.1016/j.neuroimage.2007.09.034

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 40.

  Whitfield-GabrieliSNieto-CastanonA. Conn: a functional connectivity toolbox for correlated and anticorrelated brain networks. Brain Connect (2012) 2(3):125–41. 10.1089/brain.2012.0073

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 41.

  BehzadiYRestomKLiauJLiuTT. A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage. (2007) 37(1):90–101. 10.1016/j.neuroimage.2007.04.042

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 42.

  TinazziMFiorioMFiaschiARothwellJCBhatiaKP. Sensory functions in dystonia: insights from behavioral studies. Move Disord (2009) 24:1427–36. 10.1002/mds.22490

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43.

  FiorioMTinazziMBertolasiLAgliotiSM. Temporal processing of visuotactile and tactile stimuli in writer ’ s cramp. Ann Neurol (2003) 53:630–5. 10.1002/ana.10525

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 44.

  ScontriniAConteADefazioGFiorioMFabbriniGSuppaAet alSomatosensory temporal discrimination in patients with primary focal dystonia. J Neurol Neurosurg Psychiatry (2009) 80(12):1315–9. 10.1136/jnnp.2009.178236

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 45.

  ConteADefazioGHallettMFabbriniGBerardelliA. The role of sensory information in the pathophysiology of focal dystonias. Nature reviews neurology. Nat Publishing Group (2019) 15:224–33. 10.1038/s41582-019-0137-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46.

  ElbertTCandiaVAltenmüllerERauHSterrARockstrohBet alAlteration of digital representations in somatosensory cortex in focal hand dystonia. Neuroreport (1998) 9(16):3571–5. 10.1097/00001756-199811160-00006

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47.

  JankoDThoenesKParkDWilloughbyWRHortonMBoldingM. Somatotopic mapping of the fingers in the somatosensory cortex using functional magnetic resonance imaging: a review of literature. Front Neuroanat (2022) 16:866848. 10.3389/fnana.2022.866848

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48.

  EjazNHamadaMDiedrichsenJ. Hand use predicts the structure of representations in sensorimotor cortex. Nat Neurosci (2015) 18(7):1034–40. 10.1038/nn.4038

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 49.

  MerzenichMMKaasJHWallJNelsonRJSurMFellemanD. Topographic reorganization of somatosensory cortical areas 3b and 1 in adult monkeys following restricted deafferentation. Neuroscience (1983) 8(1):33–55. 10.1016/0306-4522(83)90024-6

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 50.

  SutherlingWWLevesqueMFBaumgartnerC. Cortical sensory representation of the human hand. Neurology (1992) 42(5):1020. 10.1212/WNL.42.5.1020

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51.

  OverduinSAServosP. Distributed digit somatotopy in primary somatosensory cortex. Neuroimage (2004) 23(2):462–72. 10.1016/j.neuroimage.2004.06.024

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52.

  SchweisfurthMAFrahmJFarinaDSchweizerR. Comparison of fMRI digit representations of the dominant and non-dominant hand in the human primary somatosensory cortex. Front Hum Neurosci (2018) 12:12. 10.3389/fnhum.2018.00492

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 53.

  SathianKZangaladzeA. Tactile spatial acuity at the human fingertip and lip: bilateral symmetry and interdigit variability. Neurology (1996) 46(5):1464. 10.1212/WNL.46.5.1464

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 54.

  MohammadiBKolleweKSamiiABeckmannCFDenglerRMünteTF. Changes in resting-state brain networks in writer’s cramp. Hum Brain Mapp (2012) 33(4):840–8. 10.1002/hbm.21250

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 55.

  AsgharMSanchez-PanchueloRGloverPPakenhamDO’NeillGCSchluppeckDet alAssessing the Digit Organisation of Focal Hand Dystonia Using 7T Functional MRI (2025). 10.1101/2025.09.11.675579

  • CrossRef
  • Google Scholar