Abstract
Objective: To assess neocortical changes and their relevance to cognitive impairment in early relapsing–remitting (RR) multiple sclerosis (MS).
Methods: Conventional MR was acquired in 41 patients with RR MS and 16 demographically matched normal control subjects (NCs). An automated analysis tool was used with conventional T1-weighted MRI to obtain measures of cortical brain volumes normalized for head size. Neuropsychological performance of MS patients was assessed using the Rao Brief Repeatable Battery. Relationship between volumetric MR measures and neuropsychological scores was assessed.
Results: Neuropsychological assessment allowed for the identification of 18 cognitively preserved (MS-cp) and 23 cognitively impaired (MS-ci) MS patients. The whole MS sample showed lower values of normalized cortical volumes (NCVs) than did the NC group (p = 0.01). Upon grouping of MS patients according to cognitive performance, NCV values were lower (p = 0.02) in MS-ci patients than in both MS-cp patients and NCs. Moreover, there were positive correlations between NCV values and measures of verbal memory (r = 0.51, p = 0.02), verbal fluency (r = 0.51, p = 0.01), and attention/concentration (r = 0.65, p < 0.001) in MS-ci patients. Furthermore, NCV values were decreased in patients who scored lower on a greater number of tests (r = −0.58, p < 0.01) in the MS-ci group. None of the neuropsychological measures correlated to NCV values in the MS-cp patient group.
Conclusions: Cortical atrophy was found only in cognitively impaired patients and was significantly correlated with a poorer performance on tests of verbal memory, attention/concentration, and verbal fluency. Gray matter pathology may contribute to the development of cognitive impairment in MS from the earliest stages of the disease.
With use of computerized MR methods to measure total and regional brain volumes,1 progressive decreases in total brain volumes have been consistently reported in patients with multiple sclerosis (MS).2–7⇓⇓⇓⇓⇓ Interestingly, in MS, a primary demyelinating disease, recent postmortem8,9⇓ and in vivo10–13⇓⇓⇓ studies have also shown a selective pathology in the cerebral neocortex. Cognitive impairment can be demonstrated in 40 to 65% of MS patients,14 sometimes starting from the early stages of the disease.15,16⇓ In MS, the deficit of cognition is generally considered to be strictly dependent on white matter changes and, eventually, subcortical pathology.17–19⇓⇓ However, although increasing cognitive impairment may at times proceed in parallel with increasing T2-weighted MR lesion load, the magnitude of the correlation is generally modest.20–23⇓⇓⇓ As MRI-visible T2-weighted lesions represent only part of the MS pathology, the involvement of apparently normal brain tissue, which seems to have a great pathologic role in the disease,24 may be relevant to understanding the process underlying cognitive dysfunction in MS.23 In support of this, cognitive impairment has been associated in many MR studies with measures of cerebral atrophy,20,25,26⇓⇓ and the importance of decreasing brain volumes, rather than increasing lesion load, to MS-related cognitive impairment has been demonstrated even in patients with early relapsing–remitting (RR) MS.27,28⇓ Although a number of MS studies have associated cognitive dysfunction with MR indicators of brain tissue loss, little is known about the specific contribution of the neocortical pathology to MS-related cognitive impairment.
To assess the relevance of selective neocortical changes to cognitive impairment in early MS, we evaluated the cognitive performance as well as neocortical volumes in a cohort of RR MS patients with prevalently short disease duration and low levels of physical disability.
Methods.
Patient population.
The study sample consisted of 41 patients (30 women, 11 men) with clinically definite MS.29 Patients were recruited among those who were consecutively referred to the MS Clinics of the University of Florence and the Hospital of Empoli. Patients’ inclusion criteria were as follows: age ≤55 years (mean = 35.1 ± 8.6 years, range = 20 to 55 years, median = 34 years), RR disease course, disease duration ≤10 years (mean = 4.0 ± 2.8 years, range = 0.5 to 10 years, median = 3.5 years), and Expanded Disability Status Scale (EDSS)30 score of ≤4.0 (mean 1.5 ± 0.6, range = 1.0 to 4.0, median = 1.5). Mean educational level was 10.9 ± 3.1 years (range = 5 to 18 years, median = 13 years). All RR MS patients were relapse-free and were not taking steroids for at least 1 month before MRI and neuropsychological assessment. No patient was taking psychoactive drugs or substances that might interfere with neuropsychological performance. Two of 41 patients were being treated with interferon-β1a at the time of the study. All patients underwent identical neuropsychological and MRI protocols (see below). The study was approved by the Ethics Committee of the Faculty of Medicine of the University of Siena where MS patients underwent the MR protocol. An informed consent was obtained from all participants.
Clinical and neuropsychological assessment.
For each patient, neurologic evaluation, including disability assessment on the EDSS and neuropsychological testing, was performed within 1 week of MR examination by a neurologist and a neuropsychologist who were both blinded to the MRI results. The neuropsychological performance of MS patients was tested by using the Rao Brief Repeatable Battery (BRB),31 which incorporates tests of verbal memory acquisition and delayed recall (Selective Reminding Test and Selective Reminding Test–D), spatial memory acquisition and delayed recall (10/36 Spatial Recall Test and 10/36 Spatial Recall Test–Delayed), sustained attention, concentration, and speed of information processing (Paced Auditory Serial Addition Test at 3 and 2 seconds; Symbol Digit Modalities Test), and verbal fluency on semantic stimulus (Word List Generation). Moreover, depression was assessed through the Montgomery and Asberg Depression Rating Scale (MADRS).32 As we sought to compare neocortical brain volumes in RR MS patients with mild cognitive impairment and without even mild signs of cognitive dysfunction, the failure of at least one test on the BRB was considered as a predetermined primary measure of cognitive impairment. Therefore, we considered those patients who scored 2 SDs below the mean normative values33 on at least one test of the BRB cognitively impaired (MS-ci) and those patients who had all tests of the BRB within normal limits cognitively preserved (MS-cp).
MR examinations.
The MR protocol included a transverse dual-echo, turbo spin echo sequence (repetition time [TR]/echo time [TE]1/TE2 = 2,075/30/90 milliseconds, 256 × 256 matrix, one signal average, 250 × 250–mm field of view) yielding proton density (PD) and T2-weighted images with 50 contiguous 3-mm-thick slices, acquired parallel to the line connecting the anterior and posterior commissures. Subsequently, transverse T1-weighted gradient echo images (TR/TE = 35/10 milliseconds, 256 × 256 matrix, one signal average, 250 × 250–mm field of view) were acquired. This sequence yielded image volumes of 50 slices, 3 mm thick, oriented to match exactly the PD/T2 acquisition.
MR data analysis.
Lesion volumes.
Classification of T2-weighted lesion volume (LV) was performed for each patient by a single observer employing a user-supervised thresholding technique while being unaware of the subjects’ identity. Lesion borders were determined primarily on PD-weighted images. Information from T2- and T1-weighted images was also considered, as the software used (MEDx) offered the ability to toggle between the three sets of images, providing the operator with convenient access to the information in both data sets while defining lesions and facilitating the discrimination of CSF from periventricular plaques. Total LV was calculated by multiplying the lesion area by slice thickness. The coefficient of variation was about 5% in serial measurements.
Brain volumes.
On T1-weighted MR images, normalized volumes of the whole of the brain parenchyma and neocortical gray matter were measured using a method for total and regional brain volume measurements (the cross-sectional version of the SIENA software34 SIENAX). SIENAX uses BET (Brain Extraction Tool, part of the FSL-FMRIB Software Library; www.fmrib.ox.ac.uk/fsl) to extract the brain and skull from the MR images, as previously described.5 A tissue segmentation program (FAST, another part of FSL)35 is then used to segment the extracted brain image into gray and white matter, CSF, and background, yielding an estimate of total brain tissue volume. The brain-extracted MR images are registered on a canonical image in a standardized space (using the skull image to provide the scaling cue), a procedure that also provides a spatial normalization (scaling) factor for each subject. For selective measurements of neocortical volumes, a standard space mask (which includes ventricles, deep gray matter, cerebellum, and brainstem) is used to separate segmented gray matter into neocortical and nonneocortical. The estimated volumes for a subject are then multiplied by the normalization factor to yield either the volume of the total brain tissue (NBV) or the neocortical gray matter volume (NCV). This fully automated method provides results with an accuracy of 0.5 to 1% for single-timepoint (cross-sectional) measurements.5,34⇓
Statistical analysis.
The nonparametric Mann–Whitney test was used for comparisons of NCV values relative to the whole group of MS patients and those of demographically matched normal controls (NCs; 16 subjects; 11 women, 5 men; mean age = 36.0 ± 8 years, age range = 21 to 52 years). The NC subjects were recruited among laboratory and hospital workers and were included in the group if they had a negative history for neurologic disorders and no abnormalities on the conventional brain MRI. In subgroup analysis, values of the MS-ci and MS-cp patients were compared with those of NCs. As the subgroup of MS-ci patients was significantly older than that of MS-cp patients (table 1), before statistical comparison, MR data were corrected by using a z-score transformation relative to an age-matched NC group for each patient group (NC group for comparison with MS-ci patients = 12 subjects, mean = 38.5 ± 7, age range = 28 to 52 years; NC group for comparison with MS-cp patients = 10 subjects, mean = 31.8 ± 6, age range = 21 to 43 years). This allowed for the avoidance of potentially spurious results due to differences in age between patient subgroups. After z-score transformation, differences between patient and NC groups were assessed using analysis of variance followed by pairwise post-hoc comparison using the Tukey highest significant difference procedure to account for multiple comparisons. Relationships between MR and cognitive variables were assessed using the nonparametric Spearman rank order correlation. Data were considered significant at the 0.05 level. The SYSTAT software (version 9) running on Windows (SPSS, Chicago, IL) was used to perform statistical analysis.
Table 1 Clinical and demographic information on cognitively preserved and cognitively impaired MS patients
Results.
The neuropsychological assessment on the BRB allowed for the identification of 18 MS-cp patients and 23 MS-ci patients. Among these 23 patients, 14 patients failed only one neuropsychological test, 6 failed two, 1 failed three, and 2 patients failed four. The type of neuropsychological tests failed by the patient population and the relative percentage are reported in table 2. The MS-ci and MS-cp groups were similar for educational level, disease duration, and EDSS score, but MS-ci patients were older (p < 0.01) than MS-cp subjects (see table 1). Mean scores on the MADRS were not significantly different between the two patient groups, and with use of a score of 8 as the cut-off point,36 four subjects in the MS-cp and five subjects in the MS-ci group were classified as depressed.
Table 2 Neuropsychological tests failed by cognitively impaired MS patients
Quantitative MR analysis showed lower NCV values in the whole group of RR MS patients than in the NC group (NCV in MS patients = 603 ± 53 mL, NCV in NC = 636 ± 22 mL; p = 0.01). In contrast, the fully automated estimation of standardized NBV showed a trend toward decreased values in the MS group with respect to the age-matched NC group, but this did not reach significance (NBV in MS patients = 1,584 ± 68 mL, NBV in NCs = 1,615 ± 40 mL; p = 0.07).
When similar analyses were performed in MS patients grouped according to neuropsychological performances, after correction for common effects of age, NCV values were lower (p = 0.02) in MS-ci patients than in both MS-cp patients and NCs (NCV z-score in MS-ci = −2.025, NCV z-score in MS-cp = −1.033) (figure). In addition, MS-ci patients showed moderate to strong positive correlations between NCV values and scores on measures of verbal memory (r = 0.51, p = 0.02), verbal fluency (r = 0.51, p = 0.01), and attention/concentration (r = 0.65, p < 0.001). Furthermore, in this group of MS patients, NCV values were lower in patients who failed a greater number of tests (r = −0.58, p < 0.01). None of the neuropsychological measurements correlated to NCV values in the MS-cp patient group (data are summarized in table 3).
Figure. Box plots comparing the standardized MR measurements of normalized neocortical volume (NCV) of cognitively impaired multiple sclerosis (MS) patients (right box; MS-ci), cognitively preserved MS patients (center box; MS-cp), and normal control subjects (left box; NC). Data are z-score transformed to correct for differences in age between the two MS groups. NCV values are lower (p = 0.02) in MS-ci patients than in both MS-cp patients and NC.
Table 3 Correlations between normalized cortical volumes and neuropsychological test scores
Interestingly, MS-cp and MS-ci patients did not differ for NBV (NBV in MS-ci patients = 1,575 ± 68 mL, NBV in MS-cp = 1,595 ± 69 mL; p = 0.3) or total brain T2-weighted LV (T2-weighted LV in MS-cp = 4.1 ± 2.4 mL, T2-weighted LV in MS-ci = 5.8 ± 4.8 mL; p = 0.4) measurements. Additionally, NBV and total brain T2-weighted LV did not show any significant correlations with neuropsychological measures in the two patient groups.
Discussion.
Although several studies have investigated the possible MR correlates of cognitive dysfunction in MS, so far little is known about the specific contribution of gray matter pathology to MS-related cognitive impairment. Preliminary data on this score were provided by a recent PET study,37 in which significant cortical involvement in MS-related cognitive dysfunction has been suggested. In this study, we obtained selective measurements of neocortical volumes from T1-weighted MR images and standardized neuropsychological tests in a group of patients with RR MS. After grouping patients with mild cognitive impairment and patients with preserved cognition according to their performance on neuropsychological tests, significant decreases in NCV were selectively found in the former patient group. In addition, only MS-ci patients showed a close relationship between measures of neocortical atrophy and both global measures of the degree of cognitive impairment (expressed as the total number of tests failed by each subject) and selective measures of cognitive dysfunction such as defects in verbal and spatial memory, sustained attention and concentration, and verbal fluency. In contrast, perhaps due to the early patient disease stages, no significant differences in measures of total brain volumes and white matter T2-weighted LVs were found between MS-ci and MS-cp patients.
Although white matter pathology alone cannot explain cognitive dysfunction in MS, pathologic changes in gray matter may provide a relevant contribution. The significance of gray matter pathology in MS, although recognized by early neuropathologists,38 was not widely appreciated until post-mortem8,9⇓ and in vivo MS studies10–13⇓⇓⇓ re-examined gray matter involvement in this disease. In particular, some of these studies have shown that neocortical pathology is not only substantial in MS but can also be demonstrated in vivo in different forms of the disease starting from the earliest stages of the clinical course. Results of this study confirm these previous observations of early neocortical involvement in MS and extend them by pointing out a significant association between neocortical atrophy and cognitive dysfunction in RR MS patients. The lack of significant differences in total brain volume and in T2-weighted lesion load between our MS-cp and MS-ci patients leads to the hypothesis that the relevance of white matter pathology to MS-related cognitive impairment might be modest and that selective measures of neocortical atrophy might represent, at least at early disease stages, the most sensitive indicator of cognition integrity and dysfunction. The absence of significant correlations between T2-weighted LV and neuropsychological tests in MS-ci patients supports this hypothesis and suggests that MS-related cognitive impairment might be due, at least in part, to mechanisms that are not related to focal lesion genesis. The magnitude of the correlation between neuropsychological test scores and total brain T2-wieghted lesions is usually modest and may sometimes be greater when this is measured between regional brain lesions and specific neuropsychological tests.18,37,39⇓⇓ As we studied subjects with early disease and relatively mild brain lesion accumulation, regional measurements of T2-weighted lesions were not performed in our patients, and this might also explain the lack of correlation found in this study between T2-wieghted MR lesion load and neuropsychological measures.
The mechanisms underlying neocortical pathology in MS are not fully clarified. One hypothesis is that axonal damage may lead to retrograde neurodegeneration. As the corticospinal tracts and the frontal periventricular white matter are preferential sites for MS white matter lesions,40 white matter pathology in MS may lead to selective retrograde injury to frontal, temporal, and motor areas of the cortex. This could explain the characteristic pattern of the focal cortical thinning found in a recent study.2 However, other studies10,41⇓ suggest that retrograde changes from focal white matter lesions cannot satisfactorily explain the full range of findings. Most likely, cortical inflammatory pathology leading to myelin loss, axonal transection, and neuronal apoptosis can, in fact, lead to cortical atrophy.8,42,43⇓⇓ Experimental studies suggesting that axonal loss may occur, at least in part, via mechanisms that are independent of those causing demyelination44 and that it is probably related to the presence of an abnormal glia–axonal interaction even with low8,44⇓ or absent45 inflammation give support to the hypothesis that whereas a proportion of this neocortical pathology may be secondary to white matter inflammation, an independent neurodegenerative process also might be active. Finally, as the T2-weighted hypointense neocortical lesions show “normal” T1 signal, neocortical or juxtacortical lesions may have intensity similar to that of the gray matter and therefore could have been included in the volume measurement of the neocortex. However, as decreases in NCV have been also found in previous studies10,12⇓ even in patients with minimal white matter T2-weighted LV, it is very unlikely that the detected decrease in neocortical volumes could be due primarily to a neocortical lesion burden.
As the current version of the method used for NCV quantification (SIENAX)34 does not allow measurements in specific areas of the brain, in this study we could not assess correlations between regional neocortical atrophy and specific cognitive defects. However, in a recent study,12 cortical thinning was found at the earliest disease stages predominantly in the superior temporal gyrus and the superior and middle frontal gyri. As all these brain regions have critical importance in cognitive functioning, our findings reinforce this previous observation and suggest that specific anatomic locations could be assigned to cognitive relevant cortical changes. Longitudinal studies using quantitative approaches to imaging analysis should be carried out to better clarify these aspects as well as the time-dependent nature of cortical changes and their progressive importance on cognition in MS.
Acknowledgments
N. De Stefano was supported by the PAR grant of the University of Siena and by a grant from the Italian Society of Multiple Sclerosis.
Footnotes
-
Presented at the 55th annual meeting of the American Academy of Neurology, Honolulu, HI, April 2003.
- Received October 1, 2003.
- Accepted February 27, 2004.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
