Abstract
Background: A number of neurophysiologic and neuroimaging techniques have been evaluated in the research setting to assess upper motor neuron (UMN) damage in ALS. Changes in tissue structure in the CNS modify the diffusional behavior of water molecules, which can be detected by diffusion tensor MRI.
Objectives: To explore the hypothesis that degeneration of the motor fibers in ALS would be reflected by changes in the diffusion characteristics of the white matter fibers in the posterior limb of the internal capsule and that these changes could be detected by diffusion tensor MRI.
Methods: We studied 22 patients with El Escorial definite, probable, or possible ALS—11 with limb onset (mean age 54.5 ± 10.7 years) and 11 with bulbar onset (mean age 49.6 ± 11.7 years)—and compared them with 20 healthy, age-matched controls (mean age 46.0 ± 12.6 years). We assessed central motor conduction time (CMCT), threshold to stimulation, and silent period using transcranial magnetic stimulation. Diffusion tensor MRI was performed using a 1.5-T GE Signa system (Milwaukee, WI) fitted with Advanced NMR hardware and software capable of producing echo planar MR images. Data were acquired from seven coronal slices centered to include the posterior limb of the internal capsule. Maps of the mean diffusivity, fractional anisotropy, and T2-weighted signal intensity were generated.
Results: There were no differences between the subject groups on measures of CMCT, threshold to stimulation, and silent period. However, the CMCT correlated with clinical measures of UMN involvement. We found a significant increase in the mean diffusivity and reduction in fractional anisotropy along the corticospinal tracts between the three subject groups, most marked in the bulbar-onset group. The fractional anisotropy correlated with measures of disease severity and UMN involvement, whereas the mean diffusivity correlated with disease duration.
Conclusion: The results support the use of diffusion tensor MRI in detecting pathology of the corticospinal tracts in ALS.
The diagnosis of ALS is currently based on a history of progressive weakness with evidence of upper and lower motor neuron involvement that can only be explained based on a system disorder affecting the corticospinal tracts and somatic motor neurons. There are currently no techniques available in clinical practice for objectively assessing the upper motor neuron (UMN) damage. Such assessment has practical implications in the diagnosis of ALS in the 5 to 10% of patients in whom diagnostic doubt arises, particularly in those patients with equivocal UMN signs. It may help to determine the contribution of upper and lower motor neuron involvement to symptomatology and provide an objective marker for disease progression, particularly relevant to clinical trials. Both neurophysiologic and neuroimaging techniques have been used in the research setting to evaluate UMN pathology.
Transcranial magnetic stimulation (TMS) allows the assessment of central and peripheral motor pathway function. In ALS, central motor conduction times (CMCTs) have been measured using TMS, but results are conflicting.1-5 Although prolongation of the CMCT is reported in a proportion of patients, the measure does not provide useful data diagnostically as an indicator of disease severity or prognostically.1,6 Corticomotor threshold to stimulation may be reduced early in the disease process, but tends to be elevated when clinical signs of motor neuron degeneration are evident.7 The silent period may also vary within the disease process, with a negative linear relationship described between the silent period and disease duration.8 Both the threshold to stimulation and the silent period may offer insights into the pathophysiology of ALS, but their role in monitoring disease progression is unclear.
In neuroimaging, proton MRS (1H-MRS) has shown promise in assessing UMN pathology in the research setting.9-13 Pioro et al.9 first demonstrated a reduction in the N-acetylaspartate (NAA)/creatine (Cr) ratio in the motor cortex region in ALS. They attributed this to neuronal loss or dysfunction, or both, because NAA is found only in the nerve cell bodies and their processes. This finding has been reproduced in additional studies, both in the motor cortex region10-13 and in the brainstem,14 but there is clearly some overlap between the patients and controls.12,15 Block et al.13 carried out follow-up measurements on nine patients and demonstrated a trend toward progressive reductions in NAA/choline and NAA/Cr ratios in the motor cortex region, suggesting a role in quantifying UMN degeneration in ALS.
Changes in tissue structure within the CNS modify the diffusional behavior of water molecules, which can be detected by MRI techniques. Such methods have been used to detect early changes of ischemia before changes have been apparent on conventional MR images in both animal models16 and human cases.17 In brain tumors, diffusion measurements may be of value to assess the components of the tumor, which could help differentiate tumor subtypes.18 In studies of experimental allergic encephalomyelitis with CNS induction (an animal model of MS), despite pathologic and clinical evidence of encephalitis, myelitis, and demyelination, no abnormalities were identified on conventional T1- or T2-weighted scans during the first episode.19,20 However, lesions were seen on diffusion-weighted MR images when conventional images were normal.21 Thus, the technique is more specific to certain brain changes compared with conventional scanning procedures.
Early reports of diffusion MR measurements16,17 relied on images weighted by diffusion in a single direction, for example, parallel or perpendicular to the principle orientation of a particular nerve fiber tract. Thus, image contrast relied on both a directionally dependent measure of diffusion and factors such as the T1 and T2 of a tissue.22-24 By acquiring a series of images with differing amounts of diffusion weighting in a single direction, a quantitative measure of diffusion in one direction can be derived (the apparent diffusion coefficient). Fiber directionality means that such measurements are dependent on the direction chosen for diffusion sensitization. The advent of rapid echo planar imaging (EPI) has made it possible to acquire images with a range of diffusion weighting in multiple directions within a realistic time scale (diffusion tensor imaging).25 Not only does this allow the determination of directionally independent diffusion measures, but the series of measurements in different directions facilitates measures of the anisotropy of diffusion (i.e., the directional dependence).25-27 The mean diffusivity therefore assesses the restriction of movement of the water molecules, regardless of direction, such that a high mean diffusivity implies a less restricted environment (e.g., CSF), whereas a low mean diffusivity implies a more restricted environment. In biological systems, the water molecules encounter collisions with the cell membranes and macromolecules. Therefore, the mean diffusivity in the white matter of the brain is reduced due to collision of the water molecules with cell structures.28 However, water molecules move more readily parallel to the principal direction of structured material (in this case nerve fibers) than perpendicular to it.22-24,29 The directionality of diffusion can be quantified by an anisotropy value, which is a measure of the ratio of diffusivity along the axis of fiber bundles compared with across them. For example, the diffusion anisotropy will be higher in a highly ordered structure such as the corpus callosum than in the gray matter. Fractional anisotropy is a measure of the degree of directionality of diffusion within a single voxel resulting from alignment and ordering of tissue structures. Fractional anisotropy values range from 0 (no directional dependence of diffusion) to 1 (diffusion along a single direction).27
ALS provides an interesting candidate for diffusion tensor imaging studies because it involves degeneration of the cortico-cortical and corticospinal fibers. We studied patients with ALS and compared them with normal controls to explore the hypothesis that the degeneration of the motor fibers in ALS would be reflected by changes in the diffusion characteristics in the motor fibers of the posterior limb of the internal capsule and in the subcortical fibers in the motor cortex region.
Methods.
Clinical base.
Patients were recruited from the King’s Motor Neuron Disease (MND) Care and Research Centre, London. Twenty-two patients with El Escorial definite, probable, or possible ALS were recruited30—11 with limb onset (mean age 54.5 ± 10.7 years) and 11 with bulbar onset (mean age 49.6 ± 11.7 years)—and compared with 20 healthy, age-matched controls (mean age 46.0 ± 12.6 years). The controls were unrelated friends or spouses of the patients.
Clinical assessments.
Before inclusion in the study, all patients were assessed in the King’s MND Care and Research Centre and were recruited if they had evidence of combined upper and lower motor neuron involvement in at least one region, as defined by clinical examination supported by electrophysiologic evidence of denervation, and exclusion of other causes by appropriate blood tests and neuroimaging.
All patients and controls gave written informed consent. The study was approved by The Bethlem and Maudsley NHS Trust Research Ethics Committee.
No patient or control had a history of cerebrovascular disease, and none were taking psychoactive drugs.
All patients underwent physical examination on the day of the scanning. Disease severity was estimated using the ALS severity scale, evaluating bulbar and spinal function as well as overall disease severity.31 A low score in this scale indicates greater impairment. The scale was developed and validated by Hillel et al.,31 although it was not validated against muscle strength. We chose to use this scale because it allowed us to assess bulbar and spinal function separately. Muscle strength was assessed by the same examiner using the Medical Research Council (MRC) rating scale. An assessment of rapidity of disease progression was made as:
Rapidity = (40 − ALS severity)/duration
The modified Ashworth spasticity scale and a newly designed spasticity scale were used to assess UMN involvement. The presence of the Babinski sign, Hoffmann’s sign, and clonus were noted, and patients were categorized as having definite UMN involvement in the presence of these signs or probable UMN involvement, so-called ALS-PUMNS (probable UMN signs),9 with clear lower motor neuron signs and overactive reflexes in the same limbs but no Babinski sign, Hoffmann’s sign, or clonus. Physical assessments were completed in 30 to 60 minutes, depending on patient disability.
Transcranial magnetic stimulation.
Patients underwent TMS within 3 weeks of the scanning protocol. EMG recordings were taken from the abductor pollicis brevis muscle via surface electrodes. TMS was delivered by a Magstim 250 magnetic stimulator (The Magstim Company Limited, Spring Gardens, Whitland, Wales, UK) using a high-power 90-mm circular coil. The motor cortex was identified as the site of lowest stimulus intensity capable of inducing a consistent motor evoked potential (MEP),32 with the muscle tested in the resting state. This stimulus intensity was taken as our threshold. Four recordings were taken from the motor cortex followed by an additional four recordings with the stimulus applied over the cervical spine to calculate the CMCT. Finally, the stimulation intensity was increased to 10% above threshold, and the inhibition of ongoing voluntary muscle contraction by magnetic stimulation of the motor cortex was measured (silent period) over an acquisition of six sweeps. The absolute silent period was measured as the shortest duration over the six sweeps from the end of the MEP to the end of the absolute EMG silence.33 The protocol was performed bilaterally and took approximately 30 minutes to complete.
Diffusion tensor MRI.
A 1.5-T General Electric Signa MR system (Milwaukee, WI) fitted with hardware and software for EPI (Advanced NMR, Woburn, MA) was used for all studies. Daily quality assurance of EPI was carried out to ensure high signal-to-ghost ratio and excellent temporal stability using an automated quality control procedure.34 Head movement was limited by a vacuum fixation device. A series of images was acquired, including axial dual echo fast spin-echo images (repetition time [TR] = 4,000 msec, echo time [TE] 1 = 17 msec, TE2 = 102 msec, echo train length = 8, 22-cm field of view, 5-mm–thick slices, 0.5-mm gap, 256 × 192 acquisition matrix [frequency encoding × phase encoding], 1 data average) and coronal fast inversion recovery images optimized for contrast from previous simulations35 (TR = 3,000 msec, TI = 200 msec, TE = 15 msec, echo train length = 8, 22-cm field of view, 3.5-mm interleaved slices, 256 × 256 acquisition matrix, 2 data averages). A modified single shot spin-echo EPI pulse sequence containing two 50-msec duration magnetic field gradient pulses for diffusion weighting was applied. Data were acquired from seven coronal slices centered to include the posterior limb of the internal capsule. A field of view of 40 × 20 cm was used with an acquisition matrix of 256 × 128 (frequency encoding × phase encoding). Diffusion weighting was applied in seven noncollinear directions (x, y, z, xy, yz, xz, and xyz) with each component of the b-matrix taking eight equally spaced values between 0 and 620 sec/mm2. The total acquisition time for all the MRIs was 45 minutes.
The diffusion-weighted images were initially smoothed using a nonlinear filter to preserve information within the image while reducing the effects of noise.36 The diffusion tensor was calculated for each voxel assuming a mono-exponential relationship between signal intensity and the product of the b-matrix (a 3 × 3 matrix that characterizes the diffusion weighting in all directions) and components of the 3 × 3 diffusion tensor matrix. Multivariate regression was used to solve for all unique components of the diffusion tensor matrix together with the non–diffusion-weighted intensity.25 Properties of the diffusion tensor matrix were then computed pixel by pixel and displayed as images. The properties were the mean diffusivity, which is a measure of directionally averaged diffusion, and the fractional anisotropy.
Mean diffusivity and fractional anisotropy measurements were taken in six regions along the corticospinal tracts incorporating the posterior limb of the internal capsule, on each side, 12 regions in total (figure 1). We believed that taking the regions separately rather than tracing around the length of the tracts allowed more accurate assessment of the descending white matter fibers. The mean diffusivity and fractional anisotropy values from all 12 derived regions were averaged to provide a single mean diffusivity and fractional anisotropy value for each individual. We chose to average the values from the right and left sides because a recent publication suggested that the alignment of fiber tracts measured by diffusion-encoded MRI was asymmetric in the right and left hemispheres in the anterior limb of the internal capsule.37 Although the authors did not show asymmetry within the posterior limb of the internal capsule, we felt that averaging our results would address this possible confounding factor.
Figure 1. Regions of interest used for analysis of diffusion characteristics along white matter tracts descending through the posterior limb of the internal capsule.
The T2-weighted images corresponding to the same anatomic positions as the diffusion tensor MR images were reviewed by a consultant neuroradiologist (J.M.D.) blinded to diagnosis and diffusion characteristics for evidence of T2 hyperintensities and were categorized as having definite, probable, possible, or no such lesions.
Statistical analysis.
A one-way analysis of variance, with one between-subjects factor of group, was used to compare the two patient groups (limb-onset and bulbar-onset) or all three groups (controls, limb-onset patients, and bulbar-onset patients). In the latter, a specified contrast was also employed to compare the total ALS group and controls. Bonferroni’s correction was applied for multiple analyses. Two-tailed Pearson correlation coefficients examined hypothesis-led correlations. All statistical tests were performed at the 5% level of significance using SPSS software (Chicago, IL).
Results.
Patient characteristics (table 1). Although the controls were slightly younger than the patient groups, the difference was not statistically significant. The groups were matched for sex (p = 0.29).
Patient characteristics
On comparing the limb- and bulbar-onset patients, the total score on the ALS severity scale was lower in the bulbar-onset group (p = 0.06). The limb subscore did not differ between groups (p = 0.41), whereas the bulbar subscore was lower in the bulbar-onset group, reflecting the subdivision of patients (p < 0.00). Disease duration and rapidity of disease did not differ between groups (p > 0.14).
There were no differences between the limb- and bulbar-onset groups on measures of muscle strength (MRC grade) or spasticity, as measured by the modified Ashworth spasticity scale and the new spasticity scale (p > 0.18).
Transcranial magnetic stimulation.
Eleven limb-onset ALS patients, 6 bulbar-onset patients, and 13 controls agreed to undergo TMS. Results were discarded from one patient with limb-onset ALS and one control because recordings were inadequate.
Although the CMCTs showed a trend toward an increase in the bulbar-onset group (limb-onset group: 8.0 ± 1.1 msec; bulbar-onset group: 9.8 ± 2.8 msec; controls: 7.7 ± 1.2 msec), this did not reach statistical significance at the 5% level when the Bonferroni correction was applied. No statistically significant differences were observed between the three groups and the threshold to stimulation (limb-onset: 60.8 ± 9.4; bulbar-onset: 65.5 ± 6.9; controls: 55.9 ± 8.0) or the absolute silent period (limb-onset 77.7 ± 21.3 msec; bulbar-onset: 55.9 ± 21.5 msec; controls: 69.2 ± 20.2 msec).
No correlation was found between the measures of disease severity (ALS severity scale and MRC grades) and the threshold, CMCT, or silent period. The CMCT correlated with UMN involvement as measured by the Ashworth spasticity scale (r = 0.53, p = 0.03) and showed a trend toward correlation with the new spasticity scale (r = 0.48, p = 0.06). There was no correlation between the CMCT or threshold and the disease duration or rapidity of disease progression (p > 0.11). The silent period tended to be shorter early in the disease, but this did not reach statistical significance (r = 0.48, p = 0.08). There was no correlation between the silent period and rapidity of disease progression (p = 0.34).
Diffusion tensor MRI.
Inadequate scanning information was obtained from two patients—one limb-onset and one bulbar-onset. The results of the mean diffusivity and fractional anisotropy measurements taken from the average of the 12 regions analyzed are shown in table 2. The mean diffusivity was significantly elevated in the ALS patients compared with the controls in both the limb- and the bulbar-onset groups (figure 2). The fractional anisotropy value was reduced in the ALS group as a whole compared with controls (p = 0.02), with a significant reduction occurring in the bulbar- but not the limb-onset patients (figure 3). Individual results are shown in table 3. Nine patients had a mean diffusivity value of greater than the average mean diffusivity + 2 standard deviations (SD) in the control group (i.e., >0.78 × 10−3 mm2/sec). Six patients and one control had an average fractional anisotropy value of less than the average fractional anisotropy − 2 SD in the control group (i.e., <0.728). Four patients (three bulbar-onset and one limb-onset) had both a mean diffusivity value of greater than the average mean diffusivity + 2 SD and an average fractional anisotropy value of less than the average fractional anisotropy − 2 SD in the control group. These four patients had clear evidence of spasticity on clinical examination.
Mean ± standard deviation (SD) for the mean diffusivity and fractional anisotropy measurements from the 12 regions studied in bulbar-onset and limb-onset ALS patients and controls
Figure 2. Average mean diffusivity and 95% confidence intervals from the posterior limb of the internal capsule in limb-onset and bulbar-onset ALS patients and controls.
Figure 3. Average fractional anisotropy and 95% confidence intervals from the posterior limb of the internal capsule in limb-onset and bulbar-onset ALS patients and controls.
Individual results: Upper motor neuron (UMN) involvement, diffusion characteristics, and T2 hyperintensities
Those patients with hyperintensities visible on the T2-weighted fast spin-echo images in the posterior limb of the internal capsule are shown in table 3. None of the patients with both abnormal mean diffusivity and fractional anisotropy values had T2 hyperintensities apparent on the conventional images. One control showed possible hyperintense signal on T2-weighted images with normal diffusion characteristics.
The mean diffusivity value showed a positive correlation with disease duration (r = 0.57, p = 0.009) but did not correlate with the measures of disease severity (p > 0.29) or UMN involvement (p > 0.17). Conversely, the fractional anisotropy showed no correlation with disease duration (r = −0.17, p = 0.48) but did correlate with the ALS severity scale (r = 0.63, p = 0.003), Ashworth spasticity scale (r = −0.56, p = 0.007), and the new spasticity scale (r = −0.55, p = 0.01).
There was no correlation between the mean diffusivity or the fractional anisotropy and the CMCT or the silent period as measured by TMS (p > 0.29), although there was a correlation between both the mean diffusivity and the fractional anisotropy and the threshold to motor conduction (r = 0.60, p = 0.002 and r = −0.57, p = 0.003 respectively).
Discussion.
Using the TMS methods described, we were unable to show differences in the CMCT, threshold to stimulation, or silent period in the three subject groups. However, the CMCT correlated with clinical measures of UMN involvement in ALS and therefore may have a role in monitoring UMN pathology. We also found that the silent period tended to be shorter early in the disease process, and this may represent increased excitability or loss of inhibitory control of cortical motor neurons in the early stages of ALS.7,8
We have demonstrated an elevation in the mean diffusivity and a reduction in the fractional anisotropy in the posterior limb of the internal capsule in patients with ALS compared with normal controls. This occurred in the absence of hyperintensities on the T2-weighted images in the majority of patients. The elevation in mean diffusivity is consistent with studies on disorders leading to axonal degeneration, such as chronic MS38,39 and chronic ischemic lesions40 as well as in cerebral white matter in Alzheimer’s disease,41 all of which show an increase in the apparent diffusion coefficient in affected areas. In our patients, this may represent an increase in the extracellular volume secondary to axonal loss. A reduction in fractional anisotropy has been demonstrated in chronic MS lesions and at the core of acute MS lesions,42 therefore suggesting that this measure may be useful in assessing myelination. Both increased mean diffusivity and reduced fractional anisotropy were found in a study of leukoaraiosis.43 Changes in mean diffusivity and fractional anisotropy values are not specific to any one disease, but may add valuable information about pathology in the context of clinical findings. To our knowledge, only two studies have previously explored the use of diffusion-weighted MRI in ALS.44,45 Segawa et al.44 found no differences in diffusion characteristics in the posterior limb of the internal capsule in ALS patients compared with controls, although high signal lesions were seen on T2-weighted images in all patients. This was not the case in our patients and may suggest a different population was studied. Current technology used in the acquisition of diffusion characteristics also allows more accurate detection of changes found in pathologic states. Wu et al.45 showed diffusion-weighted hyperintensity in the corticospinal tract at the level of the internal capsule in 11 of 12 patients with ALS and in 5 of 12 controls. High signal intensity was also seen on T2-weighted images in 11 of 12 patients and 8 of 12 controls. Their method was not quantitative. To our knowledge, this is the first investigation of robust quantitative measures of mean diffusivity and fractional anisotropy in this patient population.
Although the differences were not statistically significant, our control group tended to be younger than the patient groups. Gideon et al.46 found a significant positive correlation between the apparent diffusion coefficient in cerebral white matter and age in 17 healthy volunteers. In our group of 20 controls there was no correlation between age and mean diffusivity or fractional anisotropy, and age probably does not account for the differences between our patient and control groups. Also, the most significant differences were found in the bulbar-onset group, which had a younger mean age than the limb-onset group.
We demonstrated correlations between the fractional anisotropy and measures of disease severity, rapidity of disease progression, and UMN involvement, but did not show similar correlations with the mean diffusivity. The mean diffusivity correlated with disease duration, whereas fractional anisotropy did not. This suggests that the mean diffusivity and fractional anisotropy are assessing different pathologic effects. The abnormal fractional anisotropy was found predominantly in the bulbar-onset group. We have previously shown abnormalities in the NAA/(Cr + phosphocreatine) peak area ratio using 1H-MRS in the subcortical white matter in the motor region in bulbar-onset patients compared with limb-onset patients and controls.15 These two findings may suggest a greater UMN component in the bulbar-onset group or may imply a different pathologic process in relation to the corticospinal tracts in patients with this phenotype. The CMCT also tended to be more delayed in the bulbar-onset group in this study. The corticomotor threshold to magnetic stimulation correlated with both the mean diffusivity positively and with the fractional anisotropy negatively. This is consistent with the previous finding that threshold may be reduced early in the disease but tends to be elevated when clinical signs of UMN degeneration are evident.7 Therefore the results of investigations of the central motor neurons may vary depending on both the clinical onset of disease and the stage in the disease process.
In table 3, we divided our individual patients into those with definite UMN signs—as evidenced by a Babinski sign, clonus, or positive Hoffmann’s sign—and those with probable UMN signs, so-called ALS-PUMNS, with clear lower motor neuron signs and overactive reflexes in the same limbs but no Babinski or Hoffmann’s sign or clonus. Six patients had ALS-PUMNS, but only one of these had a high mean diffusivity in the corticospinal tract. Therefore, using this methodology, diffusion tensor MRI is unlikely to be useful for confirming UMN pathology when clinical signs are equivocal.
Interestingly, in one patient with a very rapid progression of disease and severe UMN involvement clinically, clear hyperintense lesions were seen on the T2-weighted images, but the mean diffusivity was not significantly elevated compared with control values, although the fractional anisotropy was reduced. This is in keeping with our finding that the mean diffusivity value correlated with disease duration, and the pathologic process in ALS may affect the fractional anisotropy early in the disease process, whereas the elevation in mean diffusivity value represents more chronic change with loss of neurons. Interestingly, in acute ischemic lesions, the diffusivity is initially reduced, only being elevated in more subacute to chronic lesions.17 Therefore, the normal diffusivity may represent a pseudo-normalization47,48 in the transition between the initial low value in early stages of pathology and the elevated value associated with chronic axonal loss and gliosis.
The glutamate release inhibitor, riluzole, a medication licensed for the treatment of ALS, may play a role in the abnormalities shown on diffusion tensor MRI. With the patient numbers studied, we are unable to determine the relative contribution of the drug compared with the onset of disease to the changes in diffusion characteristics described. However, of the 10 patients studied in each of the limb- and bulbar-onset groups, 4 from each group were taking riluzole and 6 were not. Therefore the differences demonstrated in the fractional anisotropy between the limb- and bulbar-onset groups would be unlikely to be caused by the riluzole. The changes in diffusion characteristics secondary to riluzole require further investigation.
This study has demonstrated changes in the diffusion characteristics of the water molecules along the white matter fibers corresponding to the corticospinal tracts in patients with ALS. Despite the statistical significance, the overlap between patient and control values suggests the technique would not be of value in early diagnosis. However, longitudinal studies of diffusion tensor MRI of the corticospinal tracts in ALS are needed to assess changes in the mean diffusivity and fractional anisotropy as the disease progresses to determine its contribution to the future study of pathogenesis and drug response. Our observations suggest that diffusion tensor MRI may be useful in analyzing the extent and severity of corticospinal tract degeneration in ALS.
Acknowledgments
Supported by the Motor Neurone Disease Association UK. C.M.E. is supported by Action Research. D.K.J. was supported by the Wellcome Trust (grant 043235/Z/94/Z).
- Received November 30, 1998.
- Accepted April 10, 1999.
References
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
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.