Evidence for widespread movement-associated functional MRI changes in patients with PPMS

Patients.

We studied 30 right-handed patients with definite PPMS.¹¹ These patients were carefully selected from a large group of PPMS patients attending three MS Outpatient Clinics to exclude patients with clinical evidence of cerebellar or sensory impairments associated with the deficits of the pyramidal system. There were 18 men and 12 women; their mean age was 50.4 years (range 34 to 68 years), median disease duration was 10 years (range 2 to 28 years), and median Expanded Disability Status Scale (EDSS) score¹² was 5.5 (range 2.0 to 8.0). Patients were divided into 2 groups of 15 subjects each based on the preferential side of their motor involvement. The first group of patients (9 men and 6 women; mean age 48.9 years, range 35 to 55 years; median disease duration 12 years, range 5 to 25 years; median EDSS score 6.0, range 4.5 to 8.0) had severe paresis of the right lower limb. In four of these patients, pyramidal signs at the level of the right upper limb were also found. Five of them had also pyramidal signs of the left lower limb, whereas none of them had clinical impairment of the left upper limb. This group of patients will be referred as R-PPMS. The second group (9 men and 6 women; mean age 51.8 years, range 34 to 68 years; median disease duration 9 years, range 2 to 28 years; median EDSS score 4.5, range 2.0 to 6.0) had a mild to severe paresis of the left lower limb. In four of these patients, pyramidal signs at the level of the left upper limb were also found. None of them had any neurologic symptom or sign at the level of the right limbs. This group of patients will be referred as L-PPMS. Patients with R-PPMS had higher EDSS scores than L-PPMS patients (p < 0.001), whereas the two groups did not differ in terms of age, sex, or disease duration. Fifteen right-handed healthy volunteers with no previous history of neurologic dysfunction and a normal neurologic exam (8 men and 7 women; mean age 48.3 years, range 34 to 62 years) served as control subjects. All subjects were assessed clinically by a single neurologist who was unaware of the MRI and fMRI results. Local ethical committee approval and written informed consent from all subjects were obtained prior to study initiation.

Functional assessment.

Motor functional assessment was performed for all the subjects at the time of MRI acquisition. For the right upper limbs, the nine-hole peg test and the maximum finger-tapping frequency were used.¹³ The maximum finger-tapping rate was observed for two 30-second trial periods outside the magnet, and the mean frequency to the nearest 0.5 Hz entered the analysis. Time to complete the nine-hole peg test and finger-tapping rate were similar between patients and control subjects (table). For the right lower limbs, the maximum foot-tapping frequency was used.⁶ While foot-tapping rates were similar between L-PPMS patients and control subjects, R-PPMS patients had slower (p < 0.001) foot-tapping speeds than the previous two groups (see the table).

Experimental design.

With use of a block design (ABAB), where epochs of activation were alternated with epochs of rest, all the subjects were scanned while performing 3 different tasks, split into 3 different runs of 60 measurements each. The first task (Task 1) consisted of repetitive flexion–extension of the last four fingers of the right hand moving together. The second task (Task 2) consisted of repetitive flexion–extension of the right foot. The third task (Task 3) consisted of the performance of the first two tasks simultaneously in a phasic manner. The order of experiments was random. All tasks were paced by a metronome at 1-Hz frequency. Patients and control subjects were trained before performing the experiments. Subjects were instructed to keep their eyes closed during fMRI acquisition and were monitored visually during scanning to ensure accurate task performance and to assess for additional (e.g., mirror) movements. For clinically unimpaired limbs, tasks were performed equally well by all subjects.

fMRI acquisition.

Brain MR scans were obtained on a magnet operating at 1.5 T (Vision; Siemens, Erlangen, Germany). Sagittal T1-weighted images were obtained to define the anterior–posterior commissural (AC-PC) plane. FMR images were acquired using a T2*-weighted echo planar imaging sequence (repetition time [TR] = 96 milliseconds, echo time [TE] = 66 milliseconds, flip angle = 90°, matrix size = 128 × 128, field of view [FOV] = 256 × 256 mm, interscan interval = 5.5 seconds). Twenty-four axial slices, parallel to the AC-PC plane, with a thickness of 5 mm, covering the whole brain, were acquired during each measurement. Shimming was performed for the entire brain using an autoshim routine, which yielded satisfactory magnetic field homogeneity.

Structural MRI acquisition and analysis.

With the same magnet, a dual-echo turbo spin echo sequence (TR = 3,300 milliseconds, first echo TE = 16 milliseconds, second echo TE = 98 milliseconds, echo train length = 5, slice thickness = 5 mm, FOV = 192 × 256 mm, matrix size = 190 × 256) was also acquired from all the subjects. Volumes of lesions seen on T2-weighted scans were measured by an experienced observer, unaware of the fMRI results, using a semiautomated segmentation technique based on local thresholding.¹⁴

fMRI analysis.

fMRI data were analyzed using the statistical parametric mapping (SPM99) software.¹⁵ Prior to statistical analysis, all images were realigned to the first one to correct for subject motion, spatially normalized into a stereotaxic space,¹⁶ and smoothed with a 10-mm, three-dimensional Gaussian filter. Changes in blood oxygen level-dependent (BOLD) contrast associated with the performance of the different motor tasks were assessed on a pixel-by-pixel basis, using the general linear model¹⁵ and the theory of Gaussian fields.¹⁷ Specific effects were tested by applying appropriate linear contrasts. Significant hemodynamic changes for each contrast were assessed using t statistical parametric maps. This approach, known as fixed effect analysis, allows only assessment of the mean effect for individual subjects; the results cannot be generalized to the whole population studied. The comparisons between groups using such an approach are unsatisfactory because group differences may be heavily affected by inter-patient variability rather than reflect systematic differences between populations. Therefore, after the first-level analysis, a second-level analysis, known as random-effect analysis,¹⁸ was also performed. In other words, this two-stage approach consists in collapsing data for each subject into a single image parameterizing the effect of interest (within-subject modeling) and assessing these images across subjects using a simple between-subject model. The quantity that entered the statistical analysis was the height of the activated clusters.^15,17⇓ This is a measure of the magnitude of the BOLD changes between activation and rest on a pixel-by-pixel basis. A one-sample t-test was used to assess the activations within individual groups when performing each of the three tasks. Analysis of variance and two-sample t-test were used for comparisons between groups. For Task 1, using the same approach and linear regression analysis,¹⁸ we also evaluated the correlation between BOLD changes throughout the entire brain during task performance and T2 lesion volume. We report activations below a threshold of p < 0.05 corrected for multiple comparisons.

Conventional MRI.

All healthy volunteers had normal brain MR dual-echo scans. The mean T2 lesion volumes were 21.8 mL (range 1.1 to 75.7 mL) for the whole sample of patients with PPMS, 23.2 mL (range 4.9 to 75.7 mL) for patients with R-PPMS, and 20.3 mL (range 1.1 to 41.5 mL) for patients with L-PPMS (the difference in T2 lesion volumes between patients with R-PPMS and those with L-PPMS was not significant). T2 lesions in the motor pathways contralateral to the clinical deficits were found in six patients with R-PPMS and in four of those with L-PPMS.

Task-associated fMRI activations within individual subject groups.

The brain areas with significant activations detected while performing each of the three tasks in healthy volunteers and in patients with R-PPMS and L-PPMS, along with the corresponding Talairach coordinates of the main foci of activation and t values, are reported in tables E2 through E4 (which can be found on the Neurology Web site at www.neurology.org).

Contrasts between individual subject groups.

Task 1.

Compared with the overall PPMS sample, healthy volunteers showed more significant activation in the homolateral cerebellar hemisphere (figure 1). However, this difference was not found when comparing healthy volunteers with R-PPMS patients taken in isolation. In contrast, PPMS patients showed more significant activation (not evident in healthy volunteers at the chosen threshold) bilaterally in the superior temporal gyrus (STG) and the middle frontal gyrus (MFG) and contralaterally in the postcentral gyrus. Considering the two PPMS patient groups separately, R-PPMS patients, who were only mildly impaired on this task, showed an anterior displacement of the contralateral supplementary motor area (SMA) when compared with healthy volunteers (Talairach coordinates were −2, −22, 48 vs −6, −18, 48). This resulted in a more significant activation of the displaced contralateral SMA in R-PPMS when compared with healthy control subjects (figure 2) and patients with L-PPMS. Patients with L-PPMS showed more significant activation (not seen in the other two groups at the threshold chosen) in the homolateral insula.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Relative activation of the homolateral cerebellar hemisphere in healthy subjects vs all patients with primary progressive MS while performing Task 1 (random effect analysis, between-group comparison).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Relative activation of the contralateral supplementary motor area in patients with primary progressive MS affected on their right sides vs control subjects while performing Task 1 (random effect analysis, between-group comparison).

Task 2.

Healthy volunteers showed more significant activation in the homolateral cerebellar hemisphere and basal ganglia than PPMS patients. In normal volunteers, additional clusters of activation were identified bilaterally in the precuneus and homolaterally in the middle temporal gyrus (MTG). In PPMS patients, more significant activations were located bilaterally in the MFG and the STG (these latter clusters of activation were more prominent in R-PPMS). Although homolateral cingulate motor area (CMA) activation was not seen at within-group analysis because of the conservative threshold used, the following t values were found when applying an uncorrected p value of 0.001: healthy volunteers = 5.00, R-PPMS = 8.22, and L-PPMS = 5.77. As a consequence, R-PPMS patients, who were most impaired in performance of this task, showed relatively increased activation in the homolateral CMA when compared with healthy volunteers (figure 3) and with patients with L-PPMS. R-PPMS also had more significant activation of the homolateral postcentral gyrus relative to either of the other two groups. In L-PPMS patients, additional areas of activation were located at the level of the contralateral cerebellar hemisphere and inferior frontal gyrus (IFG). In these patients, more significant activation of the contralateral SMA was also found when compared with the other two groups.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Relative activation of the homolateral cingulate motor area in patients with primary progressive MS affected on their right sides vs control subjects while performing Task 2 (random effect analysis, between-group comparison). (A) Sagittal view; (B) coronal view.

Correlation between fMRI activation patterns and T2 lesion load.

Correlations between lesion burden and activation associated with right-hand flexion–extension (Task 1) were explored. In the R-PPMS group, which included patients who were functionally impaired, T2 lesion load was correlated with relative activation in the homolateral (r = 0.86, p < 0.001) and contralateral (r = 0.77, p < 0.001) postcentral gyri, homolateral (r = 0.84, p < 0.001) and contralateral (r = 0.81, p < 0.001) upper bank of Sylvian fissure, and contralateral calcarine sulcus (r = 0.77, p < 0.001). Even in the L-PPMS group, who showed no behavioral impairments for Task 1, there were significant changes in activation patterns with lesion load. For this group, the T2 lesion load was correlated with activation in the primary SMC bilaterally (r = 0.76, p < 0.001) (figure 4) and in the contralateral middle occipital gyrus (r = 0.70, p < 0.001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. (A) Relative activations of the homolateral primary sensorimotor cortex (SMC) and contralateral SMC and supplementary motor area (SMA) in patients with primary progressive MS affected on their left sides vs control subjects while performing Task 1 (random effect analysis, between-group comparison). (B) The correlations between homolateral primary SMC activity and T2 lesion load are shown. This correlation remained significant even after removing the two “outliers” (r = 0.70, p < 0.001).