Abstract
Objective: To compare fMRI activation during two visual stimulation paradigms in Parkinson disease (PD) subjects with chronic visual hallucinations vs PD patients who had never hallucinated.
Methods: Twelve pairs of PD subjects, matched for age, PD duration, and dopaminergic drug exposure duration, participated in this study. The authors examined group differences in activation during stroboscopic (flashing) vs no visual stimulation and kinematic (apparent motion) vs stationary visual stimulation.
Results: During stroboscopic stimulation, non-hallucinating PD subjects showed significantly greater activation in the parietal lobe and cingulate gyrus compared to hallucinating PD subjects. In contrast, the hallucinating subjects showed significantly greater activation in the inferior frontal gyrus and the caudate nucleus. During kinematic stimulation, non-hallucinating PD subjects showed significantly greater activation in area V5/MT, parietal lobe, and cingulate gyrus compared to hallucinating PD subjects. Hallucinating PD subjects showed significantly greater activation in the superior frontal gyrus.
Conclusions: PD patients with chronic visual hallucinations respond to visual stimuli with greater frontal and subcortical activation and less visual cortical activation than non-hallucinating PD subjects. Shifting visual circuitry from posterior to anterior regions associated primarily with attention processes suggests altered network organization may play a role in the pathophysiology of visual hallucinations in PD.
Parkinson disease (PD) is associated with hallucinations in approximately one third of patients on long-term dopaminergic therapy.1 In contrast to the predominantly auditory hallucinations of schizophrenia, hallucinations associated with PD are typically visual.2 Other senses may be involved in hallucinations, including tactile, olfactory, and auditory senses, but these usually accompany visual hallucinations.3 Visual hallucinations in PD are often characterized by well-formed, stereotypic, brightly colored, moving images with a duration of seconds to minutes and with retained insight.4–7⇓⇓⇓ These hallucinations primarily occur in the context of longstanding, chronically treated PD, and different studies have emphasized the role of patient age, disease duration, and treatment duration on the development of hallucinations.1,3,7⇓⇓ Studies have identified coexistent medical diagnoses that increase the risk of hallucinations, and the most consistent risk factors are poor primary vision, coexistent dementia, and coexistent depression.1,8–10⇓⇓⇓ When untreated, hallucinations persist,11 progress,12 and are a primary risk factor for nursing home placement and its related morbidities.13,14⇓
The exact pathophysiology of hallucinations in PD is unknown, but dopaminergic, and possibly serotonergic systems may play a role in their development. The anatomic substrate of hallucinations is unknown as well, and there are almost no histopathologic studies of PD patients with documented hallucinations. One autopsy study that combined subjects with Lewy body dementia, PD with dementia, and PD without dementia documented an association between the presence of visual hallucinations and Lewy bodies in the temporal lobe, specifically the amygdala and parahippocampus.15
Neuroimaging studies of PD patients with hallucinations are limited to case reports. Structural MR studies searching for occipital lobe or deep white matter lesions in the visual pathways have not detected systematic changes among hallucinators.16 fMRI offers a new technology to assess activation patterns associated with a history of hallucinations.
In this study, we used fMRI to compare cortical activation during stroboscopic and kinematic visual stimulation in PD subjects with vs without visual hallucinations. Cortical activation during stroboscopic perception assessed differences in the processing of primary visual stimulation between hallucinators and non-hallucinators. Cortical activation during kinematic perception assessed group differences in the processing of apparent movement related stimulation because the visual hallucinations in PD often involve motion.8 Of particular interest for the apparent motion condition was activation in area V5/MT in the middle temporal/parietal/occipital region, a region implicated in the perception of motion in humans.17–19⇓⇓
Methods.
Overview.
We conducted a case-control study of 12 PD patients (cases) with chronic hallucinations, matched for age, PD severity and duration, and duration of dopaminergic medication with 12 PD subjects without hallucinations (controls). All subjects underwent fMRI scans that were conducted by a technologist blind to group membership.
Subjects.
The cases were 12 patients with visual hallucinations at least three times weekly, on stable PD medications including levodopa and not on neuroleptic medications to treat hallucinations. To eliminate the possible confounding effects of dementia, we recruited only subjects with Mini-Mental State Examination20 (MMSE) scores ≥ 24 and a clinical diagnosis of no dementia. Because the fMRI protocol required full alertness and cooperation to follow visual stimuli, subjects who were claustrophobic, blind, or with clinically significant visual impairment were excluded. From our PD database and sequential chart review prior to regularly scheduled outpatient visits, we identified one control subject for each case. The 12 control participants had PD, but had never hallucinated and fulfilled the same MMSE, visual, and claustrophobia criteria used for the cases. Cases and controls were individually matched by age (±3 years), PD duration (±5 years), dopaminergic drug exposure duration (±5 years), and Hoehn and Yahr “on” stage (exact match). Visual acuity was assessed using the Snellen eye chart for both corrected and uncorrected vision. All cases and controls were right-handed and signed informed consent as approved by Rush University Office of Human Research.
Assessment of hallucinations.
Frequency of hallucinations was assessed via self-report. Severity of hallucinations was assessed using the Hallucination item from the Neuropsychiatric Inventory–Questionnaire21 (NPI-Q) for severity (scores range from 0 “Not Present” to 3 “Severe”) and distress (scores range from 0 “Not distressing at all” to 5 “Extreme or very severe”), and the Global Rating of Hallucinations item from the Scale for the Assessment of Positive Symptoms22 (SAPS) (scores range from 0 “None” to 5 “Severe”).
fMRI activation tasks.
Visual stimuli for two activation tasks were presented using a G3 Macintosh (Cupertino, CA) laptop computer running PsyScope software23 for stimuli control connected to a MRI compatible LCD projector (Resonance Technology, Van Nuys, CA). Images were back-projected onto a screen attached to the MR bore. Participants viewed the screen through a mirror mounted on the receiver head coil without correction for visual acuity.
The two activation tasks included a stroboscopic task and a kinematic task. The stroboscopic task followed that of a previously published protocol24 and was designed to assess brain activation associated with basic visual processing. During this task participants viewed alternating blocks of stroboscopic stimulation at a rate of 3 Hz and blocks of no visual stimulation. The kinematic task followed that of a previously published protocol25 and was designed to assess brain activation associated with the perception of apparent motion. During this task, participants viewed alternating blocks of stationary concentric circles and apparently moving concentric circles, radiating from the center of the screen to the edge of the screen at a rate of 1 Hz. Each task alternated in 30-second blocks for a total of 6 minutes (12 blocks). During the tasks the participants were told to watch the screen.
MRI scanning procedure.
Imaging was performed on a 1.5 T General Electric scanner (General Electric Medical Systems, Milwaukee, WI) with LX Horizon high-speed gradient upgrades (Rev 8.4) at Rush-Presbyterian-St. Luke’s Medical Center. Subjects underwent scanning within 1 hour after their usual antiparkinsonian medication dose. Head movement was minimized by using foam padding and tape to secure participants’ heads. A whole-brain T2*-weighted 2D gradient-echo spiral pulse sequence26 (repetition time [TR] = 3,000 msec; echo time [TE] = 40 msec; field of view [FOV] = 24 cm; 21 slices [thickness = 6 mm with 0 gap]; inplane resolution = 3.75 mm), which is relatively insensitive to motion artifacts due to pulsatility or minor subject movement,27 was used for functional images. A standard quadrature head coil was used for signal acquisition.
T1-weighted, flow compensated spin-warp anatomy images (TR = 500, minimum TE) were acquired for all sections that received functional scans. These images were used to correlate functional activation with anatomic structures, i.e., voxels that were found to be significantly activated during the functional scan were overlaid on these structural images.
fMRI analysis.
Image reconstruction was performed off-line by transferring the data to a Sun SparcStation (Sun Microsystems, Cupertino, CA). A gridding algorithm was employed to resample the raw data into a Cartesian matrix. Once individual images were reconstructed, all T2*-weighted images were realigned to correct for within-scan motion. The motion parameters derived from this realignment process were used as covariates of non-interest to statistically control for individual head movement. The resultant T2* volumes were then smoothed with a 8 mm full width at half maximum (FWHM) isotropic Gaussian kernel to compensate for residual between-subject variability after spatial normalization and to permit application of Gaussian random field theory to provide for corrected statistical inference.28 Time-series at each voxel were regressed on a reference waveform. The significance of this regression was assessed with a T statistic at each voxel to construct a SPM map. The reference waveform was calculated by convolving a square wave representing the time course of the alternating conditions (stroboscopic vs rest, and kinematic vs static) with an estimated hemodynamic response function template.29 To facilitate group comparisons, the structural T1-weighted volumes were spatially normalized to a standard brain template provided by SPM99 using a 12 parameter affine normalization and nonlinear adjustments with 7 × 8 × 7 basis functions.30 The spatial transformation parameters derived from normalizing the structural volume were applied to the realigned T2*-weighted images.
The comparisons of interest were stroboscopic vs no visual stimulus and kinematic vs stationary visual stimulus. Contrasts were tested with a general linear model using a random effects approach.31 This approach takes into account between-subject variability and allows generalization of inferences beyond the specific sample participants. Subject-specific contrasts, reflecting the degree of visually evoked activation, were entered into a second level analysis with one contrast per subject. This second level analysis tested for group differences in activation, and conformed to a simple analysis of variance modeling subject and group effects and a covariate (Mini-Mental State Examination [MMSE] score) to account for the potentially confounding effect of group differences in cognitive performance. Group differences in activation were assessed with the appropriate T statistic and reported at p = 0.001 (uncorrected for multiple comparisons) for completeness. However, as noted in our tables, many of our results survived a correction for multiple comparisons based upon the entire search volume. Where relevant, we also report corrections for smaller search volumes when our hypotheses were more anatomically constrained. Individual participant t-maps of the contrasts of interest were subsequently transformed to the unit normal Z-distribution to create a statistical parametric map for each contrast. We restricted our analyses of group differences to voxels that showed greater activation in the visual stimulation conditions relative to baseline conditions. This limitation is necessary because the increases in BOLD signal are thought to represent brain activation associated with visual perception, whereas decreases in signal are considered to reflect other sorts of processes.32–34⇓⇓
Results.
Subject cohort.
The 12 cases had a mean age of 71.08 (±6.39) years, a mean PD duration of 13.92 (±4.89) years, a mean dopaminergic drug exposure duration of 10.92 (±3.99) years, a mean Unified PD Rating Scale35 motor examination (UPDRS) score of 30.42 (±10.71), and a median Hoehn and Yahr “on” stage of 3 (range 2 to 4). Median Snellen visual acuity scores for right and left eyes were 50/20 and 37.5/20 for corrected vision and 100/20 and 200/20 for uncorrected vision. The controls had a mean age of 73.25 (±7.58) years, a mean PD duration of 11.17 (±3.90) years, a mean dopaminergic drug exposure duration of 8.58 (±3.63) years, a mean UPDRS score of 31.80 (±14.81), and a median Hoehn and Yahr stage of 3 (range 2 to 4). Median Snellen visual acuity scores for right and left eyes were 25/20 and 25/20 for corrected vision and 70/20 and 70/20 for uncorrected vision. Seven of the non-hallucinators and seven of the hallucinators were taking a dopamine agonist (pergolide mesylate), at equivalent dosages (mean dosage non-hallucinators = 2.52 mg/day, mean dosage hallucinators = 2.14 mg/day). For all of these measures there were no significant differences between the cases and controls. The hallucinators, however, performed slightly worse than the non-hallucinators on the MMSE, a measure of mental status (hallucinators mean MMSE = 26.17 ± 2.25; non-hallucinators mean MMSE = 27.96 ± 2.09, t22 = 2.07, p = 0.05), although no patient scored within the demented range of performance and all patients were judged to be nondemented by clinical examination. Because of this difference in MMSE scores, individual scores were entered as a covariate of no interest in the group comparisons at the second level analyses as described above.
On average, the cases experienced visual hallucinations 6.4 times per week (range 3 to 12). Their median hallucination severity score on the NPI-Q was 2.0 (range 2 to 3) and their median distress score on this measure was 1 (range 0 to 1). Their median score for the SAPS Global Hallucination item was 3 (range 2 to 4). Because the controls did not experience hallucinations, their frequency of hallucinations was 0 times per week and their scores on the NPI and SAPS scales were 0. One hallucinating participant had to be excluded from the stroboscopic analysis due to technical failure in scan acquisition.
fMRI activation to stroboscopic and kinematic stimulation for non-hallucinating and hallucinating PD patients.
Both hallucinators and non-hallucinators evidenced robust activation of occipital cortical areas (BA 17, 18, 19) during stroboscopic stimulation compared to no visual stimulation (table 1, figure 1). Non-hallucinators had additional significant activation in the left inferior parietal lobe (BA 40) and bilateral parahippocampal gyri (BA 30). In contrast, hallucinators preferentially activated the right inferior frontal gyrus (BA 44).
Table 1 Stereotactic locations and Brodmann areas (BA) of significant (p<0.001) activation during stoboscopic and apparent kinematic visual stimulation in non-hallucinating and hallucinating Parkinson disease patients
Figure 1. Representative regions of significant fMRI activation during stroboscopic vs no visual stimulation (upper panel) and apparent kinematic vs stationary visual stimulation (lower panel) between non-hallucinating Parkinson disease (PD) patients and hallucinating PD patients. Significance thresholds were set for p < 0.001 (uncorrected for multiple comparisons) for both analyses. Voxels evidencing significant activation are displayed on representative axial sections (z = z plane Talairach coordinates) on a canonical brain image. The color scale indicates the magnitude of T values with lowest appearing in dark red and the highest in bright yellow/white. The left side of the images represents the left side of the brain.
Both hallucinators and non-hallucinators evidenced significant activation of the fusiform gyrus (BA 20) and cingulate gyrus (BA 24) during apparent kinematic stimulation compared to static visual stimulation. Non-hallucinators also showed significant bilateral activation in the middle temporal/occipital lobe (BA 19), parietal lobe (BA 40, 20, 19, 7), and temporal lobe (BA 21, 22). On the other hand, hallucinators showed additional significant activation in the superior frontal lobes (BA 6).
Group differences in fMRI activation to stroboscopic and kinematic stimulation.
During stroboscopic stimulation, non-hallucinating PD patients had significantly greater activation than hallucinating PD patients in the cingulate (BA 23, 24) and inferior parietal lobe (BA 40). Hallucinating PD patients had significantly increased activation compared to the non-hallucinating PD patients in the inferior frontal lobe (BA 44) and the caudate nucleus (table 2, figure 2).
Table 2 Stereotactic locations and Brodmann areas (BA) of significant (p<0.001) differences in activation during stroboscopic and kinematic visual stimulation between non-hallucinating and hallucinating Parkinson disease patients
Figure 2. Representative regions of significant differences in fMRI activation during stroboscopic vs no visual stimulation (upper panel) and apparent kinematic vs stationary visual stimulation (lower panel) between non-hallucinating Parkinson disease (PD) patients and hallucinating PD patients. Differences were analyzed with an analysis of covariance model (ANCOVA) using Mini-Mental State Examination score as a covariate of no interest. Significance thresholds were set for p < 0.001 (uncorrected for multiple comparisons) for both analyses. Voxels evidencing significant differences between groups are displayed on representative axial sections (z = z plane Talairach coordinates) on a canonical brain image. The color scale indicates the magnitude of T values with lowest appearing in dark red and the highest in bright yellow/white. The left side of the images represents the left side of the brain. Different imaging slices are depicted to capture the areas of significant differences between hallucinators and non-hallucinators.
During apparent kinematic stimulation, non-hallucinating PD patients had significantly greater activation compared to hallucinating PD patients in the middle temporal/occipital lobe (BA 19), the supramarginal parietal lobe (BA 40), inferior parietal lobe (BA 2), and the cingulate (BA 31). Hallucinating PD patients had significantly greater activation compared to non-hallucinating patients in the superior frontal lobe (BA 6) and the caudate nucleus (see table 2, figure 2). None of these areas of significant group differences in activation during apparent kinematic stimulation overlapped with areas of significant group differences in activation during stroboscopic stimulation.
To assess group differences during apparent kinematic stimulation in area V5/MT, a region associated with the perception of motion in humans, we conducted a small volume corrected analysis using published Talairach36 coordinates for area V5/MT.17–19,25,37⇓⇓⇓⇓ Significant decreases in area V5/MT (p < 0.05 corrected at the voxel level of multiple comparisons) during apparent kinematic stimulation in hallucinators compared to non-hallucinators were replicated in these small volume corrected analyses corresponding to the differences detected in the middle temporal/occipital lobe (BA 19) in the whole-brain group comparison analyses (see table 2).
Discussion.
This study identifies cortical activation patterns associated with visual hallucinations in PD. Among hallucinating subjects, we found that visual stimuli, especially kinematic, did not activate regions of posterior cortex (occipital, temporal, parietal) involved in the perception of visual stimuli or the perception of apparent movement to the same extent as seen in non-hallucinating patients. This finding in isolation could suggest a general decrease in stimulus-dependent activation in the hallucinating subjects. However, hallucinating patients showed increased activation in anterior cortical regions, namely frontal cortex, and the caudate nucleus during visual stimulation. The finding of focal increases in cortical and subcortical activation during visual stimulation in the hallucinating patients demonstrates that the group differences are not solely due to an overall decrease of stimulus dependent activation in the hallucination group.
The group differences in cortical activation were particularly notable for the significantly increased activation in the middle temporal/occipital region in non-hallucinators compared to hallucinators during the perception of apparent motion. This region, termed V5/MT in humans, has been identified in multiple studies as responsive to the perception of apparent motion. The regions of significant group differences identified during the apparent motion condition were unique to the perception of motion and did not overlap with regional group differences documented during stroboscopic stimulation. The coordinates in our study for significant group differences in middle temporal/occipital lobe responsiveness to apparent motion correspond to the average published coordinates for region V5/MT.17–19,25,37⇓⇓⇓⇓
Visual hallucinations are common in PD, but several other neurologic disorders are also associated with similar phenomena. These diseases can affect the eye itself, brainstem, focal cortical regions, or involve a diffuse or multifocal encephalopathy. Ocular disorders or visual deprivation can cause hallucinations, the prototypic condition being the Charles Bonnet syndrome.38 For this reason, we purposefully excluded subjects who were blind or had a history of severely compromised vision. There were no significant group differences in basic visual acuity as tested by Snellen examination. Although this finding demonstrates similar levels of refractory errors between the groups, it does not account for other, more subtle visual impairments such as contrast sensitivity and hue differentiation associated with PD in general39 and in patients with PD and hallucinations specifically.40 Assessments of these possible visual function covariates on cortical activation during visual stimulation in hallucinating PD patients need to be conducted.
Brainstem disorders are classically associated with peduncular hallucinations that, like the hallucinations of PD, are vivid in quality and may cross the border between wakefulness and dreaming.9 The high prevalence of sleep disturbance and REM-related syndromes in hallucinating PD subjects suggests that brainstem dysfunction may be important to the pathophysiology of hallucinations in PD.41–43⇓⇓ We did not consider the presence of sleep disorders in the present study and the assessment of this possible covariate to cortical activation differences in hallucinating and non-hallucinating PD patients awaits further studies.
Focal cortical disease in the form of seizures or aberrant “cross talk” or “release phenomena” in association with strokes and occipital masses can be associated with repetitive, stereotypic visual hallucinations, often in a relatively hemianoptic visual field.44 Primary neurodegenerative disorders with visual hallucinations include AD and Lewy body dementia45 where diffuse cortical lesions are prominent. One autopsy study involving subjects with Lewy body dementia, PD with dementia, and PD without dementia found an association between visual hallucinations and Lewy body deposition in the amygdala and parahippocampus.15 The anatomic substrates of these varied disorders identify several structures putatively involved in the pathogenesis of visual hallucinations, but do not describe a single unifying functional circuitry.
Neuroimaging studies have investigated the underlying substrates of visual hallucinations in these conditions as well as visual hallucinations associated with PD. In subjects with visual hallucinations due to ocular disease (e.g., Charles Bonnet syndrome), but without PD, resting state SPECT measured hyperperfusion of the temporal cortex, striatum, and thalamus,46 and increased cerebral fMRI activation during visual hallucinations in the ventral extrastriate region47 has been reported. Lewy body dementia patients evidence hypoperfusion of the occipital cortex,48,49⇓ as measured by SPECT, as well as decreased occipital cortex glucose utilization, as measured by PET.50
Very few neuroimaging studies have been reported in PD subjects with hallucinations. One resting blood flow SPECT study documented significantly lower cerebral blood flow to the left temporal lobe and temporal-occipital region among PD subjects with hallucinations compared to those without hallucinations.51 These studies of visual hallucinations in PD and other disorders suggest alterations in sensory integration regions of the temporal lobe and thalamus, as well as alternation in association visual processing areas in patients with visual hallucinations due to various pathologies.
The results of this study may provide some insight into possible mechanisms involved in visual hallucinations in patients with PD. In addition to decreased cortical activation in middle temporal/occipital regions (area V5/MT), parietal lobe, and cingulate, hallucinating patients with PD had increased activation in superior and inferior frontal cortex and the caudate nucleus compared to non-hallucinating patients. Many of these regions have been implicated in attention processes involved in visual perception. Significant correlations between a test of visual attention and resting cerebral blood flow in the right occipital and right parietal lobes in non-demented, non-hallucinating patients with PD have been demonstrated.52 The specific role of posterior parietal function in attention to motion has been demonstrated in an fMRI study.37 These authors suggest that parietal function modulates the connectivity between early visual processing areas and area V5/MT. Superior frontal regions (specifically the frontal-eye-fields [BA 6]) receive input from the striatum and form reciprocal connections to the parietal lobe and the prefrontal cortex and may further mediate visual attention.53 The alterations of parietal and frontal cortical activation during visual stimulation suggest a disruption to attentional modulation of visual perception in hallucinating patients with PD.
We conceptualize that normal visual perceptions are externally driven, “bottom-up” information processes, whereas visual hallucinations can be considered as internally driven, “top-down” processes. The hallucinating PD patients demonstrated a pattern of activation that could predispose them to hallucinations with decreased responsiveness to external perceptions in posterior cortical areas and aberrant frontal activation that gives rise to sensory visual experience. The decreased occipital registration among hallucinators suggests a role of decreased visual input or visual attention. The aberrant frontal signaling, however, suggests the possibility of primary cortical dysfunction and a misalignment of occipital-frontal circuits. Disruption of visual signal input, through decreased interplexiform retinal dopamine54 along with decreased striatonigral input,55 may lead to a loss of signal synchrony at the cortical level.56 The consequent weakening of retinal-striatal-cortical signals could lead to disinhibition of “top-down” cortical processing and aberrant release of previously stored schemas in the form of internal images.57,58⇓ The disinhibition of “top-down” signaling would overly activate visual areas that typically give rise to phenomenology interpreted as visual perceptions of the external world. These results, therefore, suggest an alteration in functional brain relations that could predispose individuals to hallucinations. Newer functional imaging techniques such as dynamic casual modeling59 may help determine cause and effect relationships between occipital, temporal, and frontal functions in this process.
Our sample is relatively small, and although carefully matched with control subjects without hallucinations, larger studies are needed. Further experiments can include a study of hallucinating subjects before and after neuroleptic treatment. Whereas these agents are highly effective in abating hallucinations,60 in our experience, treatment does not resolve hallucinations completely. Because hallucinations can fluctuate over time,12 a sequential study of patients with hallucinations during a period when they have few and then frequent hallucinations would allow for a longitudinal analysis of hallucinatory states. Ideally, scanning patients when they are having a hallucination and then repeating it when they are not hallucinating would be the most direct study. In the present study the patients were not hallucinating during the scanning session, so the difference in activation patterns did not delineate regions involved in hallucinatory experiences per se. Rather, the difference in patterns reflects the cortical activation in response to low-level visual stimulation that is associated with the chronic hallucinatory state. Our own attempts to precipitate hallucinations in an experimental setting, even with high dose levodopa infusions, have failed,61 suggesting that the heightened attention and vigilance associated with an experimental setting may suppress hallucinations and make this type of experiment difficult to achieve in a scanner. Contributing factors that could rectify the functional differences between hallucinators and non-hallucinators will be important to identify and may contribute to future therapeutic interventions.
Acknowledgments
Supported by a center grant from the PD Foundation, NY.
- Received January 19, 2004.
- Accepted June 17, 2004.
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.