Investigations on maladaptive plasticity in the sensorimotor cortex of unilateral upper limb CRPS I patients

Abstract

Background:

Patients with a complex regional pain syndrome (CRPS) in the upper limb show a sensory and motor impairment of the hand. Decreased intra-cortical-inhibition (ICI) of the motor representation of the affected hand muscle and decreased somatosensory hand representation size were related to maladaptive plasticity.

Objective:

To achieve new insights about CRPS we examined whether these alterations were present in a single cohort.

Methods:

We used a multi-modal approach comprising behavioral testing, transcranial magnetic stimulation, and high resolution fMRI combined with a new analysis technique for improved neuronal specificity.

Results:

We found a decreased pinch-grip performance, two-point discrimination on the fingertips, ICI in the motor cortex, and representation size of the hand in Brodmann Area 3b (BA3b) in the somatosensory cortex. Our analysis further showed that correlations with ICI on the non-affected side were absent on the affected side.

Conclusions:

This study is the first to gather behavioral, neurophysiologic and imaging measurements for one patient cohort and it therefore enables a comprehensive view of collapsed associations of function and representation focused on the hemisphere contralateral to the affected hand.

Keywords

Complex regional pain syndrome neuropathic pain sensorimotor cortex short-latency intra cortical inhibition high resolution fMRI

1. Introduction

In a healthy cortex, behavioral parameters such as tactile spatial resolution and characteristics of cortical excitability and representation are closely associated (e.g., Höffken et al., 2007). In patients suffering from chronic pain these associations are disrupted, resulting in what are described as maladaptive processes (Flor, 2008; Woolf & Salter, 2000). This maladaptive plasticity alters the affective, sensory, and motor system of the brain (Apkarian & Flor, 2015). Maladaptive processes are thought to be a major driver of chronic pain and their identification appears to be of central importance in the development of interventions for chronic pain patients (Lotze & Moseley, 2015). However, associations of functional parameters (e.g., TMS-and fMRI-parameters) and performance measures (e.g., somatosensory discrimination thresholds) may differ distinctively between patients suffering from chronic neuropathic pain and those suffering from complex regional pain syndrome, for instance(CRPS; Kuttikat et al., 2016). This indicates the need to identify alterations which are typical for CRPS and consistent across cohorts.

CRPS is associated with persistent pain, somatosensory and motor impairments as well as decreased usage of the affected limbs (Harden et al., 2010a). In particular, reduced somatosensory and motor performance of the affected limbs are typical clinical findings in these patients. fMRI investigations showed a smaller hand size representation within the primary somatosensory cortex on the affected side (Pietro et al., 2015) as well as a smaller intra cortical inhibition of the primary motor cortex (Strauss et al., 2015). However, similar alterations were found in healthy participants due to cast immobilization (Lissek et al., 2009; Zanette et al., 2004), raising the question as to whether pain-induced immobilization might cause the observed alterations.

In order to achieve new insights about the consistency of these alterations across cohorts we investigated: 1) Behavioral impairments by examining pinch-grip performance (PG), somatosensory two-point-discrimination (TPD) and cutaneous sensory thresholds (CST). 2) Neurophysiological alterations by measuring short-latency intra-cortical-inhibition (SICI) of the first dorsal interosseous muscle (FDI) in M1 using double pulse transcranial magnetic stimulation (TMS). 3) Hand representation size in Brodmann area (BA) 3b using fMRI. Here we tested the performance and clinical parameters, TMS assessments of cortical excitability, and high resolution fMRI-mapping of contralateral S1 BA 3b in the same CRPS-patient cohort for the first time.

2. Material and methods

2.1. Participants

The 10 patients (4 female; age: 58±8 years (mean±s.d.)) in our cohort (Table 1) were diagnosed as having CRPS I according to the Budapest criteria (Harden et al., 2010a), all suffer from a fracture of the wrist, and unilateral CRPS I affecting one of their hands with a mild sensory and motor impairment. All participants were right handed (Oldfield, 1971) and four were affected on the right hand. The symptoms were on average present for 8±12 months and clinical symptoms were scored with the “Disabilities of the Arm, Shoulder and Hand (DASH)-score” (Gummesson et al., 2003), CRPS Severity Score (CSS) (Harden et al., 2010b), and with a visual analogue score (VAS; 100 mm) for average pain rating. All participants gave their written informed consent, approved by the national ethics committee of the German Society for Psychology (ML 04_2013). For association analyses between fMRI-activation in S1 and behavioral parameters (pain: r = 0.69, two-point resolution; r = 0.77) having at least 9 patients should provide statistically relevant effects (G ^*Power 3.1; alpha 0.05; error probability 0.95) based on prior observations (Pleger et al., 2006b).

Table 1
Demographic and clinical data of the patients

Patient No Age¹ Gender² Handed ness³ Affected limb⁴ Inciting event⁵ current pain (VAS) average pain (VAS) CSS⁶ DASH⁷ Time from onset (months) Therapy⁸

1 76 M 95 L RF 0 1 7 1.7 72 A

2 59 F 85 L RF 2.4 4.3 15 50 22 B, D, E

3 55 F 95 R OC 3 5.4 15 64.7 8 –

4 39 F 88 R RF 2.3 2.4 16 70.7 8 A, C, E

5 59 M 100 R RF 4.8 7.2 7 69 5 A

6 64 M 100 L RF 2.4 2.3 12 40.5 17 A

7 55 F 100 R RF 2 2.7 6 50.9 7 –

8 63 M 100 L RF 1.9 4.7 15 57.1 11 A, B, E

9 57 M 95 L RF 2.1 4.5 15 52.4 4 A, E

10 42 M 96 L UF 1 2.3 6 14.6 3 A, E

Average 58 4F 96 R4 2.2 3.5 14 52 8

IQR 8 6 0.5 2.4 8 24 12

Patient No	Age¹	Gender²	Handed ness³	Affected limb⁴	Inciting event⁵	current pain (VAS)	average pain (VAS)	CSS⁶	DASH⁷	Time from onset (months)	Therapy⁸
1	76	M	95	L	RF	0	1	7	1.7	72	A
2	59	F	85	L	RF	2.4	4.3	15	50	22	B, D, E
3	55	F	95	R	OC	3	5.4	15	64.7	8	–
4	39	F	88	R	RF	2.3	2.4	16	70.7	8	A, C, E
5	59	M	100	R	RF	4.8	7.2	7	69	5	A
6	64	M	100	L	RF	2.4	2.3	12	40.5	17	A
7	55	F	100	R	RF	2	2.7	6	50.9	7	–
8	63	M	100	L	RF	1.9	4.7	15	57.1	11	A, B, E
9	57	M	95	L	RF	2.1	4.5	15	52.4	4	A, E
10	42	M	96	L	UF	1	2.3	6	14.6	3	A, E
Average	58	4F	96	R4		2.2	3.5	14	52	8
IQR	8		6			0.5	2.4	8	24	12

¹Age in years. ²sex: m: male, f: female. ³Edinburg handedness score (all strongly right handed). ⁴L: left, R: right. ⁵RF = radius fracture; OC = olecranon fracture; UF = ulnar fracture. ⁶CSS: CRPS severity score. ⁷DASH: Disability of Arm, Shoulder and Hand Score. ⁸Medication and Therapy: A: NSAID, non-steroidal anti-inflammatory drugs; B: opioid; C: amitriptyline; pregabalin; E: occupational-therapy.

In previous investigations, patients comparable to the ones in our cohort exhibited unilateral alterations on the affected side in: 1) Their behavioral test performance. 2) Their intra-cortical-inhibition (ICI) in the primary motor cortex (M1) investigated with double pulsed transcranial magnetic stimulation (TMS) (Eisenberg et al., 2005; Strauss et al., 2015). 3) Their hand representation size in the somatosensory cortex (S1) investigated using MEG (Juottonen et al., 2002; Maihofner et al., 2003) and fMRI (Pietro et al., 2015; Pleger et al., 2006a, 2004).

2.2. Behavioral testing

Behavioral testing was carried out immediately preceding the TMS examination. Cutaneous sensory thresholds (CST; Fig. 1 A) were investigated on the tip of the first (D1) and fifth (D5) finger on both hands (von Freyhair filaments, Touch-TestTM Sensory Evaluator, North Coast Medical, Inc., Morgan Hill, CA, USA). Two-point discrimination (TPD; Fig. 1 B) was determined on D1 using a wheel-discriminator (Sensidisk, Hannover, Germany). TPD was tested in a decreasing order of space intervals from 15 mm – 1 mm. For each interval our standardized measurement protocol included pseudo randomized sequences of filament and wheel stimuli. Pinch-Grip (PG) performance was assessed using the Roeder Manipulative Aptitude Test (Lafayette Instrument Company, Lafayette, IN, USA; Fig. 1 C). This entailed measuring, for both hands, the time needed to screw small rods into two rows, each with ten holes.

Fig.1

Behavioral, TMS and fMRI investigations. (A) Cutaneous sensory thresholds (CST) assessment (von Freyhair). (B) Two point discrimination (TPD) determination with a wheel discriminator. (C) Pinch grip performance for sensorimotor assessment (Roeder). (D) In fMRI examinations tactile stimuli were applied to the tip of D1 and D5 via pneumatic finger clips. (E) Different MEPs after single pulse non conditioned (NC) and double-pulse conditioned (C) stimulation and calculation of percentual short intracortical inhibition (% SICI). (F) Time sequence of active and rest blocks as well as the pneumatic stimulation pulses within the active blocks illustrated in a bar code scheme.

2.3. MRI data acquisition

Imaging was carried out with a 3 Tesla MRI scanner (Verio, Siemens, Germany) using a 32 channel head coil. For fMRI examinations, we applied a standard gradient echo planar imaging (EPI) sequence: resolution 1.5×1.5×2 mm³, 17 slices parallel to the post central gyrus, field of view (FoV) = 191×179 mm², echo time (TE) / repetition time (TR) = 36/2000 ms, flip angle = 76°, partial Fourier factor = 6/8, total acquisition time (TA) = 12:14 min. The first seven scans were discarded in order to achieve stable properties in the remaining fMRI time series. Characteristics of the structural scans were: Multi-Echo-MPRAGE sequence optimized for examination of the cortex (van der Kouwe et al., 2008); isotropic resolution 1 mm³, 176 sagittal slices, FoV = 256×256 mm², TE1/TE2/TE3/TE4/TR = 2.14/4/5.86/7.72/2530 ms, inversion time (TI) = 1100 ms, flip angle = 7°, echo spacing 9.8 ms, PAT GRAPPA PE 2, TA = 6:03 min.

2.4. fMRI paradigm

Stimulus application and display of the paradigm were electronically controlled with Presentation (Neurobehavioral Systems Inc., Albany, USA). Pneumatic stimulus finger clips (MEG International Services Ltd., Coquitlam, Canada) were used for tactile stimulation (Fig. 1 D). We stimulated low threshold mechanoreceptors (Merkel disks) with a frequency of 3 Hz (McGlone & Reilly, 2010). In the fMRI examinations D1 and D5 on both hands were stimulated on the distal phalanx. The fingers of each hand were examined in a single session using a block design. Stimulation and rest blocks had a duration of 10 s and 14 s respectively and were repeated 30 times (Fig. 1 F).

In each active block either D1 or D5 was stimulated, in an alternating sequence. The average stimulation frequency within the active blocks was 3 Hz. The start of the stimulation pulses was randomly varied during the active blocks and had a length of 50 ms. In each active block up to five randomly distributed pulses with a longer duration (150 ms) were inserted for temporal identification to maintain participants’ attention throughout the paradigm (Pfannmoeller et al., 2016a; Schweisfurth et al., 2011). Participants were instructed to count the total number of long pulses and to report them after each session. The amplitude of short and long pulses was equal and was regulated by the air pressure supplied to the finger clips (2 bars).

2.5. fMRI Image processing and evaluation of cortical maps

The spatial specificity of BOLD-fMRI with GE-EPI sequences is limited by artifacts from draining veins at the pial surface (Menon, 2012) which cause large displacements of the BOLD signal from the site of active neurons (Turner, 2002). Fortunately, the influence of the draining veins is mostly limited to the upper layers of the cortex (Polimeni et al., 2010). In 7 Tesla investigations a separation of the cortex in a superficial, middle and a deep layer demonstrated that the fMRI maps in middle and deep layers agree with the expected cortical maps and must therefore be free of pial vein artifacts (Ahveninen et al., 2016; Nasr et al., 2016). In 3 Tesla fMRI the voxel sizes are larger (lower limits: 7 Tesla 0.75 mm ⟶ 3 Tesla 1.5 mm) and in the range of the cortical thickness within BA3b of 1.8 mm (Meyer et al., 1996). Therefore, it is more challenging to achieve a sampling in the lower half of the cortex. Blurring due to smoothing caused by interpolations during data analysis (e.g. motion correction and registration) decreases the spatial specificity of the extracted maps even if there are no draining vein artifacts at all (Pfannmoeller et al., 2016a, 2016b). If draining vein artifacts are present the blurring enhances their range towards the lower layers of the cortex. Thus, a removal of the blurring constitutes a double gain in spatial specificity and was implemented in our data analysis to achieve a time series sampling of the BOLD signal in the lower half of the cortex.

In the implementation we applied FreeSurfer 5.3 (Fischl, 2012) for reconstruction of the white matter surface from the T1-weighted brain scan (Ecker et al., 2010). The analysis was restricted to the BA3b region between thumb (D1) and little finger (D5), including an enlargement of 4 standard deviations to account for inter individual variability (Weibull et al., 2008) in the contralateral hemisphere (Fig. 2 A - D). Boundary based register (BBR) (Greve & Fischl, 2009) was used to register the gray/white boundary in each EPI volume to the reconstructed white matter surface in the cortical BA3b hand region (BA3b-D1D5) leading to a total of 360 registrations for each hand. After registration, we inspected the positioning of all slices visually and determined the motion parameter. The volume fraction values (i.e. contributions from gray matter and CSF) were calculated for each voxel using the FreeSurfer tool mri_compute_volume_fractions. Subsequently we sampled the fMRI contrasts and the volume fraction values at the white matter surface, a quarter of the cortical thickness, and half the cortical thickness using the FreeSurfer tool mri_vol2surf and a nearest neighbor interpolator. The information on the volume fraction was used to reconstruct an fMRI time series with minimized contribution from the pial surface veins and maximized gray matter contribution (Fig. 2 E) by choosing an appropriate cortical depth at each time point. After construction of the time series for surface based analysis the FS-Fast (Tsao et al., 2003) standard methodology was used to compute general linear model (GLM) contrasts (stimulation against rest), including movement regressors. The peak value approach was used to remove residual draining vein artifacts from the surface maps (Schweisfurth et al., 2011) indicated by similar peak voxel positions for D1 and D5 stimulation. Finally, the hand size was determined as the surface distance between D1 and D5 peak voxel positions using the Dijkstra algorithm (Dijkstra, 1959).

Fig.2

Position of the BA3b hand region and sampling of fMRI time series: (A) Grey matter surface of a right hemisphere, positions of primary somatosensory cortex (S1) and hand knob (Ω) are indicated. (B) Inflated gray matter surface including the position of S1, Ω and BA3b. (C) Flattened cortex showing the BA3b-D1D5 hand region. (D) Enlargement of the hand region showing an exemplary D1D5-distance determination. (E) 2-D projection of the scheme for fMRI time series sampling assuming a flat cortex and head motion. Gray matter is indicated by gray shading and cortex borders by full lines above the rectangular grid indicating fMRI voxels. Dashed lines indicate the surface normal along the white matter surface grid points. The position at which the time series was sampled is indicated by black dots and the sampled voxels in magenta.

2.6. TMS-SICI measurements

In our SICI examinations we investigated the difference between the motor-evoked potentials (MEPs) of the first dorsal interosseous muscle (FDI) in response to single and double TMS pulses, applied to the motor cortex. The single pulses are non-conditioned pulses (NC) consisting only of the TMS test stimulus and the double pulses are conditioned pulses (C) encompassing a conditioning pulse which precedes the test stimulus by 2 ms (Fig. 1 E) (Peurala et al., 2008). SICI was calculated as a ratio of the mean MEPs evoked after NC relative to C pulse stimulation and has been related to intra-cortical-inhibition (ICI) in the motor cortex (Kujirai et al., 1993; Ziemann et al., 1996). The SICI protocol consisted of 17 stimulations including 5 NC and 12 C pulses presented in a pseudo-randomized order, comparable to previous studies (Petoe et al., 2013; Strauss et al., 2015). The MEPs were determined using electromyography (EMG) of the electric potential generated by muscle activity in the FDI.

We also measured resting motor threshold (RMT, lowest TMS intensity evoking MEPs≥50μV in relaxed FDI), active motor threshold (AMT, lowest TMS intensity evoking MEPs≥100μV over active FDI) and a recruitment curve with continuously increasing TMS output intensity. RMT served as a measure for the difference in the response between the patients and hemispheres and was necessary to adjust the individual intensity of TMS output determining the recruitment curve. Test stimulus intensity for NC and C pulses was set to be on the lower linear ascending part of the recruitment curve. The stimulus intensity of the conditioning pulse in the C pulses was set to 80% of the AMT. The pre-trigger root mean square EMG (rmsEMG) was computed over the period 30 – 100 ms prior to the test stimulus. Trials with a pre-trigger rmsEMG activity > 10μV were excluded from analysis.

EMG signals were amplified (CED 1902; Cambridge Electronic Design, Cambridge, United Kingdom), filtered (20–1000 Hz) and stored for offline analysis (Signal V.08). Single and double TMS pulses were produced by two Magstim BiStim200 stimulators (Magstim Company, Dyfed, UK). TMS was delivered using a figure-of-eight coil (see Fig. 1 E) that was held tangentially to the scalp above the area that was identified as producing the maximal MEP for the target muscle during stimulation (Petoe et al., 2013). TMS was performed on both sides of the brain to compare hemispheres. In order to avoid pharmacological effects on SICI, anti-convulsive medication was paused for at least 24 hours before the experimental procedures started. TMS and fMRI investigations were carried out less than 3 hours apart.

2.7. Statistical analysis

Distributions were tested for normality using the Lilliefors test revealing that several data sets did not have a normal distribution. We therefore applied non-parametric statistical methods except for the comparisons with values from the literature in the discussion where we used a parametric Welch test. This was convenient due to the fact that the literature results were given as mean and standard deviation and valid since our results were normally distributed in those cases. The average values and their variation were determined using the median and inter-quartile-range (IQR). Results were compared using Wilcoxon signed rank tests and Cliff’s delta as the effect size (where | Cliff’s Δ| indicates a small effect size around 0.15, a medium effect size around 0.3, and a large effect size around 0.5). Pearson correlations were used to investigate correlations between our examination results (since they have also been applied in previous investigations used for our power analysis: Pleger et al., 2006b). The correlation coefficient ρ (corresponding effect size), and linear regression was used for visualization (always valid since residual errors were normally distributed). In order to correct for possible influence of hand dominance significant results from any group analyses between the left and the right hand comparison were not used for the affected/non-affected comparison. We further tested all results for confounding by age, gender, and handedness using a X²-test but found no evidence for statistical dependencies. In all tests the FDR-adjusted threshold was used to correct for multiple comparisons, where the largest total number of tests was 15 in the comparison of the examination results for each side (Section 2.2). The statistical evaluation was performed using R (Statworx, Frankfurt/Wien/Zürich).

3. Results

3.1. Comparison of affected and unaffected sides

The affected and unaffected hands differed in PG performance (Roeder test), TPD as well as the BA3b D1D5-distances (see Table 2, Fig. 3). In addition, there was a tendency for a side difference in terms of SICI (see Table 2, Fig. 3) while the left and right hand sides revealed no differences at all (p_adj = n.s.; |Cliff’s Δ|≤0.42). We did not find any significant difference in baseline physiological parameters (i.e. RMT, AMT, MEP amplitude for NC stimulus) between the affected and unaffected sides (p_adj = n.s.; |Cliff’s Δ|≤0.27) or left and right hand sides (p_adj = n.s.; |Cliff’s Δ|≤0.42).

Table 2
Comparison of average and inter-quartile-range (IQR) for the affected and unaffected side together with the results of Wilcoxon signed rank tests (Cliff’s delta) for side differences as well as the effect size

Affected Unaffected P¹ p_adj² Cliff’s Δ

D1D5-distances¹ 17.1±3.6 21±4.9 0.006 0.01 – 0.5

TPD-D1² 5.6±2.5 3.5±2.2 0.004 0.01 0.7

SICI³ 27.5±37.1 49.9±21.7 0.04 0.06 – 0.6

Roeder 2.2±0.8 1.3±0.14 0.002 0.01 0.7

CST-D1⁴ 0.16±0.0 0.06±0.12 0.4 0.4 0.3

CST-D⁵ 0.16±0.1 0.12±0.1 0.1 0.2 0.2

	Affected	Unaffected	P¹	p_adj²	Cliff’s Δ
D1D5-distances¹	17.1±3.6	21±4.9	0.006	0.01	– 0.5
TPD-D1²	5.6±2.5	3.5±2.2	0.004	0.01	0.7
SICI³	27.5±37.1	49.9±21.7	0.04	0.06	– 0.6
Roeder	2.2±0.8	1.3±0.14	0.002	0.01	0.7
CST-D1⁴	0.16±0.0	0.06±0.12	0.4	0.4	0.3
CST-D⁵	0.16±0.1	0.12±0.1	0.1	0.2	0.2

¹p-value (uncorrected). ²p-value adjusted (corrected). ³D1-D5: thumb to pinky. ⁴TPD: two point discrimination. ⁵SICI: short intracortical inhibition. ⁶CST: cutaneous sensory thresholds.

Fig.3

Depiction of the distributions for the affected and unaffected side. (A) Pinch grip (PG) performance as assessed with the Roeder-Test. (B) SICI as assessed for the FDI. (C) Two point discrimination (TPD) of D1. (D) distance between the thumb and the pinky (D1-D5) in BA3b. (E) Cutaneous sensory thresholds (CST) for D1. (F) CST for D5. The connections between the sides indicate either a FDR significant result (^*p_adj≤0.05) or a FDR trend (^**p_adj≤0.06) for a side difference.

3.2. Associations for affected and unaffected sides

On the affected side only the correlation between D1D5-distances and CST-D1 was significant (ρ= 0.79; p_adj = 0.04; corrected for handedness; Fig. 4 A) but no other associations. For the unaffected side, we found three significant correlations which all involved SICI (Fig. 4 B): SICI correlated negatively with PG performance (Roeder test; ρ= – 0.73; p_adj = 0.04), negatively with TPD-D1 (ρ= – 0.75; p_adj = 0.04) and positively with D1D5-distances in BA3b (ρ= 0.71; p_adj = 0.04). The correlations between SICI and D1D5-distance for the affected and unaffected sides are shown in Fig. 5.

Fig.4

Schematic illustration of correlations between physiological and performance measures for (A) the affected and (B) the unaffected side. Arrows indicate significant correlations for each hemisphere.

Fig.5

Correlation between SICI and D1D5-distances on (A) the affected and (B) the unaffected side including the Pearson correlation coefficient ρ, FDR corrected probability p_adj and the result of a linear regression fit.

4. Discussion

4.1. Alterations between the affected and unaffected side

We found a decreased PG performance in the Roeder test, a decrease in TPD, and decreased BA3b hand size on the affected side, in line with previous examinations (Harden et al., 2010a; Di Pietro et al., 2015; Strauss et al., 2015). In contrast, the CSTs showed no side differences, most probably because CSTs depend predominantly on processes in the peripheral nerves (Kandel et al., 2013) and CRPS I patients per definition show no nerve damage. Furthermore, we found that the ICI of the unaffected hemisphere was in the normal range of healthy participants (SICI_healthy = 50 %; p_Welch = 0.34; t_Welch = 0.98) (Opie & Semmler, 2014; Petoe et al., 2013). In contrast ICI decreased on the affected hemisphere to about half of that value (p_Welch = 0.04; t_Welch = 2.2) indicating a difference on the affected side when compared to healthy controls. However, there was only a tendency for a difference in ICI between the affected and unaffected sides in patients. This likely indicates that some influence of hand dominance remains since the dominant hand was always the right hand while the affected hand in our patients was either the left or the right hand. The side difference in the BA3b hand size (2.9±2.4 mm) was smaller than has previously been reported for S1 (5±2.1 mm) (Pietro et al., 2015) (p_Welch = 0.05; t_Welch = 2.1). This could be a consequence of methodological improvements made to our investigation such as the inclusion of cytoarchitecture (analysis in BA3b instead of S1), surface distance measurements instead of Euclidean distances, and a higher specificity to gray matter due to the smaller fMRI voxel size and improved data analysis. Thus, our results demonstrate that these alterations occur consistently within a single cohort and therefore seem to be typical for CRPS.

However, immobilization of the upper limb is known to cause comparable alterations in M1 (Zanette et al., 2004) and S1 (Lissek et al., 2009) as observed for CRPS patients. Since chronic pain also causes an immobilization of the affected limb additional measures are necessary to determine whether the observed alterations are due to immobilization or maladaptive plasticity. Correlations between our examinations differ for the affected and non-affected sides; on the non-affected side this appears to be a correlate of everyday usage (motor performance is associated with SICI), but these associations broke down on the affected side. There is no obvious reason why reduced usage should cause a disruption of these correlations - rather a modification should be expected. It is more reasonable to assume that pain-related neuroinflammation drives this disruption, for instance due to alterations of cortical excitability (Walker et al., 2014). Thus, disrupted correlations could signal maladaptive plasticity and would in this case be an important indicator allowing differentiation between adaptive and maladaptive plasticity. We are only starting to understand maladaptive (inhibition of plastic processes, dystonia, spasticity, neuropathic pain) and pro-adaptive mechanisms but a major indication seems to be a breakdown of the function-performance associations which can be observed in the healthy brain.

4.2. Limitations of the study

In order to draw final conclusions on the correlation-based analysis, we need to know more about correlations in healthy controls and patients having a comparable fracture but no CRPS. The examination of healthy controls would provide confirmation that the correlations exist on both hands, and the examination of patients with a fracture is important to clarify how cast immobilization affects the correlations. In addition, the number of C (5 per participant) was quite low and more C stimuli might be better suited to achieve more robust TMS-results. Furthermore, non-parametric statistical approaches for correlation analyses would increase the reliability of the testing and larger samples are certainly necessary to confirm our findings. In our argumentation we make use of the findings of other studies that used healthy control groups for our comparisons with findings observed in CRPS-patients. This is problematic since the paradigms and methods differ between the studies.

All of our patients were right-handed, did not receive training to be converted from left-to right-handedness, and therefore had the same hand dominance (Grabowska et al., 2012). However, the affected hand was not identical across the group which may obscure the results due to a residual influence of hand dominance in our breakdown into affected and unaffected hand. Therefore, it would be better to investigate cohorts in which hand dominance and the affected hand is the same for all patients in order to increase the homogeneity and clarity of the results.

5. Conclusions

Our results show that the alterations between the affected and non-affected side of CRPS I patients in behavioral outcomes, M1-ICI, and S1 hand size are consistently found within a single cohort and are seemingly typical for CRPS patients. The correlations of the examination results may enable differentiation between maladaptive alterations and pain-related plastic changes, e.g. due to immobilization. In addition, our examinations were enabled by a new preprocessing for high resolution fMRI data, which is presented here, and shows the importance of an increased neuronal specificity in the investigation of chronic pain patients.

Funding

The work was not supported by other funding organizations besides the University Medicine of Greifswald.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Footnotes

Acknowledgments

Dr. Renate Schweizer and Dr. Amanda Kaas for valuable discussion and help in the data interpretation. Dr. Shahin Nasr and Dr. Avery Berman for proof reading.

References

Ahveninen, J., Chang, W.-T., Huang, S., Keil, B., Kopco, N., Rossi, S., Bonmassar, G., Witzel, T., & Polimeni, J.R. (2016). Intracortical depth analyses of frequency-sensitive regions of human auditory cortex using 7T fMRI. Neuroimage, 143, 116–127.

Apkarian, A.V., & Flor, H. (2015). Functional reorganization of the brain with chronic pain, in: The Brain Adapting with Pain. Wolters Kluwer Health Adis (ESP).

Ball, T., Breckel, T.P.K., Mutschler, I., Aertsen, A., SchulzeBonhage, A., Hennig, J., & Speck, O. (2012). Variability of fMRI-response patterns at different spatial observation scales. Human Brain Mapping, 33, 1155–1171.

Dijkstra, E.W. (1959). A note on two problems in connexion with graphs. Numerische Mathematik, 1, 269–271.

Ecker, C., Marquand, A., Mourao-Miranda, J., Johnston Daly, E.M., Brammer, M.J., Maltezos, S., Mur-.M., Robertson, D., Williams, S.C., & Murphy, .G.M. (2010). Describing the Brain in Autism in FiveeDimensions-Magnetic Resonance Imaging-Assisted Diagnosis of Autism Spectrum Disorder Using a Multiparameter Classification Approach. Journal of Neuroscience, 30, 10612–10623.

Eisenberg, E., Chistyakov, A.V., Yudashkin, M., Kaplan, B., Hafner, H., & Feinsod, M. (2005). Evidence for cortical hyperexcitability of the affected limb representation area in CRPS: A psychophysical and transcranial magnetic stimulation study. Pain, 113, 99–105.

Fischl, B. (2012). FreeSurfer. Neuroimage, 62, 774–781.

Flor, H. (2008). Maladaptive plasticity, memory for pain and phantom limb pain: Review and suggestions for new therapies. Expert Review of Neurotherapeutics, 8, 809–818.

Grabowska, A., Gut, M., Binder, M., Forsberg, L., Rymarczyk, K., & Urbanik, A. (2012). Switching handedness: fMRI study of hand motor control in right-handers, left-handers and converted left-handers. Acta Neurobiologica Experimentalis (Warsz.), 72, 439–451.

10.

Greve, D.N., & Fischl, B. (2009). Accurate and robust brain image alignment using boundary-based registration. Neuroimage, 48, 63–72.

11.

Gummesson, C., Atroshi, I., & Ekdahl, C. (2003). The disabilities of the arm, shoulder and hand (DASH) outcome questionnaire: Longitudinal construct validity and measuring self-rated health change after surgery. Bmc Musculoskeletal Disorders, 4, 11.

12.

Harden, R.N., Bruehl, S., Perez, R.S.G.M., Birklein, F., Marinus, J., Maihofner, C., Lubenow, T., Buvanendran, A., Mackey, S., Graciosa, J., Mogilevski, M., Ramsden, C., Chont, M., & Vatine, J.-J. (2010a). Validation of proposed diagnostic criteria (the “Budapest Criteria”) for Complex Regional Pain Syndrome. Pain, 150, 268–274.

13.

Harden, R.N., Bruehl, S., Perez, R.S.G.M., Birklein, F., Marinus, J., Maihofner, C., Lubenow, T., Buvanendran, A., Mackey, S., Graciosa, J., Mogilevski, M., Ramsden, C., Schlereth, T., Chont, M., & Vatine, J.-J. (2010b). Development of a severity score for CRPS. Pain, 151, 870–876.

14.

Hoffken, O., Veit, M., Knossalla, F., Lissek, S., Bliem, B., Ragert, P., Dinse, H.R., & Tegenthoff, M. (2007). Sustained increase of somatosensory cortex excitability by tactile coactivation studied by paired median nerve stimulation in humans correlates with perceptual gain. Journal of Physiology, 584, 463–471.

15.

Juottonen, K., Gockel, M., Silen, T., Hurri, H., Hari, R., & Forss, N. (2002). Altered central sensorimotor processing in patients with complex regional pain syndrome. Pain, 98, 315–323.

16.

Kandel, E.R., Jessell, T.M., Schwartz, J.H., Siegelbaum, S.A., & Hudspeth, A.J. (2013). Principles of Neural Science, Fifth Edition. McGraw-Hill Education.

17.

Kujirai, T., Caramia, M., Rothwell, J., Day, B., Thompson, P., Ferbert, A., Wroe, S., Asselman, P., & Marsden, C. (1993). Corticocortical inhibition in human motor cortex. Journal of Physiology-London, 471, 501–519.

18.

Kuttikat, A., Noreika, V., Shenker, N., Chennu, S., Bekinschtein, T., & Brown, C.A. (2016). Neurocognitive and Neuroplastic Mechanisms of Novel Clinical Signs in CRPS. Frontiers Human Neuroscience, 10, 16.

19.

Lissek, S., Wilimzig, C., Stude, P., Pleger, B., Kalisch, T., Maier, C., Peters, S.A., Nicolas, V., Tegenthoff, M., & Dinse, H.R. (2009). Immobilization Impairs Tactile Perception and Shrinks Somatosensory Cortical Maps. Current Biology, 19, 837–842.

20.

Lotze, M., & Moseley, G.L. (2015). Theoretical Considerations for Chronic Pain Rehabilitation. Physical Therapy, 95, 1316–1320.

21.

Maihofner, C., Handwerker, H.O., Neundorfer, B., & Birklein, F. (2003). Patterns of cortical reorganization in complex regional pain syndrome. Neurology, 61, 1707–1715.

22.

McGlone, F., & Reilly, D. (2010). The cutaneous sensory system. Neurosci Biobehav Rev, 34, 148–159.

23.

Menon, R.S. (2012). The great brain versus vein debate. Neuroimage, 62, 970–974.

24.

Meyer, J.R., Roychowdhury, S., Russell, E.J., Callahan, C., Gitel-man, D., & Mesulam, M.M. (1996). Location of the central sulcus via cortical thickness of the precentral and postcentral gyri on MR. Am J Neuroradiol, 17, 1699–1706.

25.

Nasr, S., Polimeni, J.R., & Tootell, R.B.H. (2016). Inter-digitated Color-and Disparity-Selective Columns within Human Visual Cortical Areas V2 and V3. J. Neurosci, 36, 1841–1857.

26.

Oldfield, R. (1971). The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia, 9, 97–113.

27.

Opie, G.M., & Semmler, J.G. (2014). Age-related Differences in Short-and Long-interval Intracortical Inhibition in a Human Hand Muscle. Brain Stimulation, 7, 665–672.

28.

Petoe, M.A., Jaque, F.A.M., Byblow, W.D., & Stinear, C.M. (2013). Cutaneous anesthesia of the forearm enhances sensorimotor function of the hand. Journal of Neurophysiology, 109, 1091–1096.

29.

Peurala, S.H., Mueller-Dahlhaus, J.F.M., Arai, N., & Ziemann, U. (2008). Interference of short-interval intracortical inhibition (SICI) and short-interval intracortical facilitation (SICF). Clinical Neurophysiology, 119, 2291–2297.

30.

Pfannmoeller, J., Greiner, M., Balasubramanian, M., & Lotze, M. (2016a). High-resolution fMRI investigations of the fingertip somatotopy and variability in BA3b and BA1 of the primary somatosensory cortex. Neuroscience, 339, 667–677.

31.

Pfannmoeller, J., Schweizer, R., & Lotze, M. (2016b). Automated Analysis Protocol for High Resolution BOLD-fMRI Mapping of the Fingertip Somatotopy in Brodmann Area 3b. J Magnetic Resonance Imaging, 43, 479–486.

32.

diPietro, F.D., Stanton, T.R., Moseley, G.L., Lotze, M., & McAuley, J.H. (2015). Interhemispheric somatosensory differences in chronic pain reflect abnormality of the Healthy side. Human Brain Mapping, 36, 508–518.

33.

Pleger, B., Blankenburg, F., Bestmann, S., Ruff, C.C., Wiech, K., Stephan, K.E., Friston, K.J., & Dolan, R.J. (2006a). Repetitive transcranial magnetic stimulation-induced changes in sensorimotor coupling parallel improvements of somatosensation in humans. Journal of Neuroscience, 26, 1945–1952.

34.

Pleger, B., Ragert, P., Schwenkreis, P., Förster, A.F., Wilimzig, C., Dinse, H., Nicolas, V., Maier, C., & Tegenthoff, M. (2006b). Patterns of cortical reorganization parallel impaired tactile discrimination and pain intensity in complex regional pain syndrome. Neuroimage, 32, 503–510.

35.

Pleger, B., Tegenthoff, M., Schwenkreis, P., Janssen, F., Ragert, P., Dinse, H.R., Volker, B., Zenz, M., & Maier, C. (2004). Meanesustained pain levels are linked to hemispherical side-to-side differences of primary somatosensory cortex in the complex regional pain syndrome I. Experimental Brain Research, 155, 115–119.

36.

Polimeni, J.R., Renvall, V., Zaretskaya, N., & Fischl, B. (2018). Analysis strategies for high-resolution UHF-fMRI data. Neuroimage, 168, 296–320.

37.

Schweisfurth, M.A., Schweizer, R., & Frahm, J. (2011). Functional MRI indicates consistent intra-digit topographic maps in the little but not the index finger within the human primary somatosensory cortex. Neuroimage, 56, 2138–2143.

38.

Schweizer, R., Voit, D., & Frahm, J. (2008). Finger representations in human primary somatosensory cortex as revealed by highresolution functional MRI of tactile stimulation. Neuroimage, 42, 28–35.

39.

Strauss, S., Grothe, M., Usichenko, T., Neumann, N., Byblow, W.D., & Lotze, M. (2015). Inhibition of the primary sen-sorimotor cortex by topical anesthesia of the forearm in patients with complex regional pain syndrome. Pain, 156, 2556–2561.

40.

Tsao, D.Y., Freiwald, W.A., Knutsen, T.A., Mandeville, J.B., & Tootell, R.B.H. (2003). Faces and objects in macaque cerebral cortex. Nature Neuroscience, 6, 989–995.

41.

Turner, R. (2002). How much cortex can a vein drain? Downstream dilution of activation-related cerebral blood oxygenation changes. Neuroimage, 16, 1062–1067.

42.

van der Kouwe, A.J.W., Benner, T., Salat, D.H., & Fischl, B. (2008). Brain morphometry with multiecho MPRAGE. Neuroimage, 40, 559–569.

43.

Walker, A.K., Kavelaars, A., Heijnen, C.J., & Dantzer, R. (2014). Neuroinflammation and Comorbidity of Pain and Depression. Pharmacological Review, 66, 80–101.

44.

Weibull, A., Bjorkman, A., Hall, H., Rosen, B., Lundborg, G., & Svensson, J. (2008). Optimizing the mapping of finger areas in primary somatosensory cortex using functional MRI. Magnetic Resonance Imaging, 26, 1342–1351.

45.

Woolf, C.J., & Salter, M.W. (2000). Neuronal Plasticity: Increasing the Gain in Pain. Science, 288, 1765.

46.

Zanette, G., Manganotti, P., Fiaschi, A., & Tamburin, S. (2004). Modulation of motor cortex excitability after upper limb immobilization. Clinical Neurophysiology, 115, 1264–1275.

47.

Ziemann, U., Rothwell, J.C., & Ridding, M.C. (1996). Interaction between intracortical inhibition and facilitation in human motor cortex. Journal of Physiology-London, 496, 873–881.