Change in PASAT performance correlates with change in P3 ERP amplitude over a 12-month period in multiple sclerosis patients.

Abstract

To examine the correlation between the change in PASAT and the change in P3 event-related potentials (ERPs) over a 12-month period in multiple sclerosis (MS) patients, and to compare the 12-month change in the P3 ERP between MS patients and controls.
Forty-four subjects (27 MS patients, 17 controls) completed visual and auditory two-stimulus oddball and three-stimulus oddball tasks at Month 0 and Month 12. Data were recorded from a 128-scalp channel electroencephalography array. Data from scalp channels were converted into continuous interpolated images (incorporating the entire scalp and time). Amplitude, topographical differences and correlations were then tested using statistical parametric mapping.
The change in visual and auditory P3a correlated significantly with the change in PASAT score (r=0.56, p<0.001 and r=0.48, p=0.003, respectively). Visual P3b and P3a showed greater decrease in 12 months in MS patients relative to controls. Visual P3b, auditory P3b and auditory P3a amplitudes had significantly decreased in MS patients after 12-month period.
Change in visual and auditory P3a ERP amplitudes correlate with change in PASAT scores in MS patients. Visual modality is more sensitive to changes in P3 ERP amplitudes over 12-month period.
P3 ERPs may have utility in monitoring the change in cognitive functioning in MS.

Full text

Page 1

Change in PASAT performance correlates with change in P3 ERP amplitude over a
12-month period in multiple sclerosis patients
H. Kiiskia,1, R.B. Reillya,1, R. Lonerganb, S. Kellyb, M. O'Brienc, K. Kinsellab, J. Bramhamc, T. Burkec,
S. O'Donnchadhac, H. Nolana, M. Hutchinsonb, N. Tubridyb, R. Whelana,⁎
aTrinity Centre for Bioengineering, Trinity College Dublin, College Green, Dublin 2, Ireland
bDepartment of Neurology, St. Vincent's University Hospital, Elm park, Dublin 4, Ireland
cSchool of Psychology, University College Dublin, Ireland
a b s t r a c t a r t i c l e i n f o
Article history:
Received 6 December 2010
Received in revised form 28 February 2011
Accepted 4 March 2011
Available online 31 March 2011
Keywords:
Multiple sclerosis
Longitudinal
Auditory event-related potentials
Visual event-related potentials
Cognition
Neuropsychology
Objective: To examine the correlation between the change in PASAT and the change in P3 event-related
potentials (ERPs) over a 12-month period in multiple sclerosis (MS) patients, and to compare the 12-month
change in the P3 ERP between MS patients and controls.
Methods: Forty-four subjects (27 MS patients, 17 controls) completed visual and auditory two-stimulus oddball
and three-stimulus oddball tasks at Month 0 and Month 12. Data were recorded from a 128-scalp channel
electroencephalography array. Data from scalp channels were converted into continuous interpolated images
(incorporatingtheentirescalpandtime).Amplitude,topographicaldifferencesandcorrelationswerethentested
using statistical parametric mapping.
Results:The change in visual and auditory P3a correlated significantly with the change in PASAT score (r=0.56,
pb0.001 and r=0.48, p=0.003, respectively). Visual P3b and P3a showed greater decrease in 12 months in MS
patients relativetocontrols. VisualP3b, auditory P3b and auditoryP3aamplitudeshad significantly decreased in
MS patients after 12-month period.
Conclusions: Change in visual and auditory P3a ERP amplitudes correlate with change in PASAT scores in MS
patients. Visual modality is more sensitive to changes in P3 ERP amplitudes over 12-month period.
Significance: P3 ERPs may have utility in monitoring the change in cognitive functioning in MS.
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
1.1. Assessing cognitive impairment in multiple sclerosis
Cognitive impairment (CI) affects up to 65% of multiple sclerosis
(MS) patients and can occur in the absence of physical disability [1].
Dysfunctions in speed of information processing, attention, memory
and executive functions are most typically observed in MS patients
[2], and CI has an impact on daily life [3,4]. It is important to recognise
CI as early as possible and to monitor its course frequently [5]. Disease
duration and disability level are usually poorly correlated with CI in
MS [5]. Neuropsychological tests in the MS Functional Composite [6],
such as the Paced Auditory Serial Addition Test (PASAT; a difficult test
of attention and working memory), are generally used to measure CI
in MS. The PASAT is influenced by practice effects [7], education,
anxiety and physical ability which is usually impaired in MS [8], and
thus cognitive electrophysiological measures have potential to assess
CI in MS.
1.2. Previous electrophysiological studies in MS
Several studies have investigated the relationship of CI and event-
related potentials (ERPs) by employing a two-stimulus oddball task, in
which occasional target stimuli have to be detected in a train of
frequent irrelevant standard stimuli. A P3b component is typically
elicited approximately 300 ms after each target stimulus, with
maximal amplitude over the parietal scalp area. The P3b is considered
to reflect context updating [9] and/or the categorization of task
relevant events [10]. Prior cross-sectional studies have reported that
PASAT score correlates with auditory P3b latency [11], and with both
visual and auditory P3b amplitudes in progressive MS patients [12].
Furthermore, the previous cross-sectional electrophysiological stud-
ies have reported MS patients to have decreased amplitudes and
increased latencies in auditory P3b [11,13–18] and in visual P3b
[11,14,16] compared to controls. Some studies [18,19] have observed
no differences between MS patients and controls.
TheP3ahasalsobeenstudiedinMS[18,20,21].TheP3aisthoughtto
be a reflection of the engagement of attention mechanisms [9] and it is
elicited by a three-stimulus oddball paradigm: in addition to standard
and target stimuli additional, infrequent, non-target stimuli are
presented. There are a variety of different types of P3a: the difficulty
Journal of the Neurological Sciences 305 (2011) 45–52
⁎ Corresponding author. Tel.: +353 1 8964214; fax: +353 1 679 5554.
E-mail address: robert.whelan@tcd.ie (R. Whelan).
1These authors contributed equally to this study.
0022-510X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jns.2011.03.018
Contents lists available at ScienceDirect
Journal of the Neurological Sciences
journal homepage: www.elsevier.com/locate/jns

Page 2

ofthestandard/targetdiscriminationandtheperceptualdistinctiveness
of the non-target stimulus alter the latency, amplitude and topography
of the P3a. Previous cross-sectional studies have reported the PASAT
scoretobecorrelatedwithvisualP3aamplitude[21]andbothvisualand
auditory P3a amplitudes in progressive MS patients [12]. Moreover,
prior studies have reported reduced auditory P3a amplitude [20], and
decreased visual P3a amplitude and delayed visual and auditory P3a
latencies relative to controls [21], whereas some [18] did not find any
differences.
1.3. Longitudinal electrophysiological studies on CI in MS
There has been some interest in studying longitudinally a variety
of evoked potentials to better monitor the progression of disability in
MS [22,23]. To our knowledge, the relationship between the PASAT
and P3 ERPs in MS has not been studied longitudinally, but a few
studies [19,24–26] have examined the change in P3b ERPs longitu-
dinally in MS. These studies have focused on the effects of interferon
treatment on cognition. Flechter et al. [24] studied 16 MS patients
using an auditory P3b paradigm prior to and one year after they
received interferon β-1b treatment. They reported that auditory P3b
amplitude decreased and latencies were shortened one year later, but
theydid notinclude a controlgroupin thestudy. Studiesthatincluded
a control group failed to find any longitudinal differences in visual or
auditory P3b after interferon treatment [19,25,26]. These studies
showed improved performance in neuropsychological measures of
attention and memory [19,24–26] after the treatment period.
Most of the previous studies of P3 ERPs in MS [19,24,26] employed
an auditory two-stimulus oddball paradigm. However, some previous
studies [11,12,21] have shown that differences between patients and
controls were influenced by the modality of the task and the P3 task
type (i.e., either two- or three-stimulus oddball). Moreover, all
longitudinal P3 ERP studies to date have focused on investigating
the effectsof interferon treatmentoncognition in relapsing–remitting
MS (RRMS) patients, and thus there is no comprehensive longitudinal
information on CI and P3 ERPs in a larger MS population, which
includes secondary-progressive MS (SPMS) patients. Another issue is
the lack of control group in the only study [24] reporting statistically
significant differences in auditory P3b over one-year period. Including
a control group in the longitudinal studies is important as P3 ERPs are
demonstrated to change with age [27] and therefore the changes seen
in the P3 components might be due to age, not disease progression or
treatment. Previous longitudinal studies employed low-density EEG
(i.e. 3–4 electrodes). In the present study a high-density EEG array
wasemployed andstatisticalparametricmapping(SPM8)wasusedto
statistically test the correlations and differences of the data. Previous
studies [11,12,21] have applied the SPM method to cross-sectional
EEG data, but SPM has not yet been employed in longitudinal studies.
This approach has the benefit that differences in scalp EEG activity
across space and time can be tested statistically.
1.4. Hypotheses
The aims of the present study were to examine the relationship
between the change in PASAT score and the change in amplitude and
topographyinP3ERPsover12-monthperiod,andtocomparethechange
betweenMSpatientsandcontrolsinthevisualandauditoryP3ERPtasks
by employing a high-density EEG array. Based on earlier studies [11,21]
we hypothesized that 1) over a 12-month period, a change in PASAT
score would correlate with change in P3 amplitude. Furthermore, we
hypothesized2)MSpatientstohavegreaterdecreaseinamplitudesover
a12-monthperiodrelativetocontrols,andthat3)decreaseinamplitude
would be best detected by visual P3b and P3a tasks.
2. Methods
2.1. Subjects
Twenty-sevenpatients(meanage43.64 yearsatMonth0)satisfying
the revised McDonald criteria for MS [28] were recruited. The MS
patients were age-matched with 17 healthy volunteers. Exclusion
criteria included current use of benzodiazepines or neuroleptics, a
historyofalcoholordrugmisuse,headinjury,orstroke.ForMSpatients,
the mean years of symptomswas 16.26 years (SD 11.60), and the mean
years of disease was 11.05 years (SD 9.64) at Month 0. One MS patient
and two controls were left-handed. Table 1 displays the demographic
data of the subjects at Month 0 and Month 12. The median number of
days between the two ERP testing sessions was 374. Two patients and
two controls were unable to complete some of the P3 ERP tasks and
three controls and two patients did not complete the PASAT in either
Month0orMonth12(seeTable1).Tothebestofourknowledge,theMS
patients did not take part in any other studies including the PASAT, or
Table 1
Demographical and behavioural data of subjects. PASAT = Paced Auditory Serial Addition Test, EDSS = Expanded Disability Status Scale, RT = reaction time, SD = standard
deviation, IQR = interquartile range. *The correlation between the 12-month PASAT difference score and 12-month P3 ERP amplitude change.
MS (total N=27) Control (total N=17)
Year 1 Year 2Year 1 Year 2
Male/female
Auditory P3a & PASAT*, (RRMS, SPMS)
Visual P3a & PASAT*, (RRMS, SPMS)
Auditory P3a, (RRMS, SPMS)
Visual P3a, (RRMS, SPMS)
Auditory P3b, (RRMS, SPMS)
Visual P3b, (RRMS, SPMS)
12/15
23 (12, 11)
23 (12, 11)
25 (14, 11)
25 (14, 11)
26 (15, 11)
26 (15, 11)
Mean
5/12
13
14
16
17
16
17
MeanSD MeanSDSDMean SD
Age
PASAT (%)
Auditory P3b RT (b800 ms)
Visual P3b RT (b800 ms)
43.64
73.59
367.30
369.92
Median
11.21
20.39
69.17
71.33
IQR
44.61
75.13
369.60
389.51
Median
11.11
21.88
58.73
73.73
IQR
40.30
85.00
369.37
366.89
Median
9.53
13.67
13.67
51.24
IQR
41.75
88.10
392.87
394.59
Median
9.29
9.36
79.02
64.47
IQR
EDSS
Hamilton
Interferon β-1a
Interferon β-1b
Natalizumab
Fingolimod
No current treatment
2.50
4.00
7
8
5
1
6
4.00
7.00
3.00
5.00
6
7
7
1
6
5.00
9.75
N/A
.00
N/A
N/A
N/A
N/A
N/A
N/A
5.00
N/A
N/A
N/A
N/A
N/A
N/A
2.00
N/A
N/A
N/A
N/A
N/A
N/A
3.00
N/A
N/A
N/A
N/A
N/A
46
H. Kiiski et al. / Journal of the Neurological Sciences 305 (2011) 45–52

Page 3

otherwiseperform thePASAT,duringthe12-monthperiod.Onepatient
had two relapses and another patient one relapse during the 12-month
period. Ethical approval was obtained from the St. Vincent's University
Hospital Ethics Committee. Informed consent was obtained from all
subjects at both time points.
2.2. Procedure
Subjects completed the standard 3-s Paced Auditory Serial
Addition Test PASAT [29] at both time points. The subjects sat with
the examiner in a quiet room, and were asked to add consecutive
single-digit numbers as they were presented on a compact disk and to
respond orally with the accurate sum. The standard PASAT form,
consisting of 61 single digits with a 3 s inter-stimulus interval, was
used. PASAT score was based on the total number of correct responses
from a maximum of 60 correct answers. Subjects were asked to
perform calculations silently, without writing or using fingers, and a
practicesequence was administered prior to the test. The subjects also
completed a measure of physical disability in MS, Kurtzke Expanded
Disability Status Scale (EDSS) [30], and the 21-item Hamilton rating
scale for depression (HRSD-21) [31].
The ActiveTwo Biosemi™ electrode system from 134 electrodes
(128 scalp electrodes), organized according to the 10–5 system [32]
and digitized at 512 Hz, was used to record ERP data in a
soundproofed room. The vertical and horizontal electro-oculograms
were recorded bilaterally from approximately 3 cm below the eye and
from the outer canthi respectively. The visual P3b paradigm consisted
of blue circles, separated by an inter-stimulus interval of 2 s,
presented for 205 trials in a pseudorandom order. Frequent non-
target (80%) and infrequent target (20%) circles were 2 cm or 4 cm in
diameter, respectively. The auditory P3b paradigm consisted of tones,
separated by an inter-stimulus interval of 2 s, presented binaurally for
205 trials in a pseudorandom order. Frequent non-target (80%) and
infrequent target (20%) tones were presented at 500 Hz and 1000 Hz
respectively. Subjects were instructed to press a button as quickly as
possible following a target stimulus. The visual P3a paradigm
consisted of stimuli separated by an inter-stimulus interval of 2 s,
presented for 410 trials in a pseudorandom order across two separate
runs of 205 trials each. Frequent standard (80%) and infrequent target
(10.24%) circles were 3.5 cm or 4 cm in diameter, respectively. The
non-target stimulus was a checkerboard (9.76%) which was 5.25 cm
per side. The auditory P3a paradigm consisted of stimuli separated by
an inter-stimulus interval of 2 s, presented binaurally for 410 trials in
a pseudorandom order across two separate runs of 205 trials each.
Frequent standard (80%) and infrequent target (10.24%) tones were
presented at 900 Hz and 1000 Hz respectively. The non-target (9.76%)
stimulus was a white noise burst. Subjects were instructed to press a
button as quickly as possible following a target stimulus. Order of
modality and task were counterbalanced across subjects. The same
procedure was repeated at Month 12.
2.3. Data analysis
Reaction times were truncated at 800 ms in order to reduce the
effects of outliers [33]. EEGLAB [34] was used for all ERP analyses. The
EEG data were bandpass filtered between 1 and 90 Hz, bandstopped
between 48 and 52 Hz, average referenced across all scalp electrodes
(appropriate when using a high-density EEG array), epoched and
baseline corrected (200 ms before baseline). Epochs with large,
obvious artefacts (e.g., muscle twitch) were first rejected manually.
Independent components analysis, using the Infomax algorithm, was
used to identify artifacts, which were subsequently removed (see
[35]). Ocular artifacts were removed by identifying the components
that correlated most highly with the EOG channels (minimum
correlation of 0.5). Visual inspection of the EOG channel before and
after removal of thecomponentwasperformedin order to ensurethat
the ocular artifacts were removed. A 99% confidence interval was
calculated across all channels for mean amplitude and variance: any
channel falling outside the confidence interval was interpolated.
Statistical analyses on demographical and descriptive data were
completed using PASW Statistics 18.0 (Predictive Analytics SoftWare;
SPSS Inc., Chicago, IL, USA, www.spss.com). SPM8 (Statistical Parametric
Mapping 8, http://www.fil.ion.ucl.ac.uk/spm) is a statistical software
package that can be used to analyze high-density EEG data. SPM8 was
used to test for topographical differences in ERP amplitude across the
entire scalp and across time. Data from each subject were transformed
intotwo-dimensionalsensor-space(interpolatedfrom128channelsand
smoothed,full-widthhalf-maximum6:6:4 a.u.)andpost-stimulus times
were set to 230–600 ms for responses in the P3b tasks, and 200–600 ms
for responses in the P3a tasks (voxel size=2.1 mm×2.7 mm×3.9 ms).
This transformation produced a three-dimensional spatiotemporal
characterization of the ERP, and these images were then used in the
subsequent analyses. The significance level was set to pb0.005 and this
was further restrained by onlyretaining scalpareas in which atleast500
contiguous voxels were statistically significant.
12-monthdifference images of each subject (showingthe topograph-
ical difference in ERP amplitude between Month 0 and Month 12 of one
subject across the entire scalp and across time) were created using the
SPM8 Image Calculator by subtracting Month 12 image from Month 0
image of the same P3 ERP task. These 12-month difference images were
thencorrelatedwiththe12-monthdifferencePASATscores(computedby
subtracting PASAT score (%) at Month 12 from PASAT score at Month 0).
A paired t-test design was employed and each individual subject's
Month0imagewasmatchedwiththecorrespondingMonth12imageof
the same P3 ERP tasks. The 12-month difference images were also
compared between MS patients and controls, to show the difference in
the 12-month change of the topography of ERP amplitude between the
two groups. Based on our previous studies we restrained the spatial
characterization of activations with spatial masks: to a large parietal
area for the P3b tasks (as reported in [11]), and to an area covering a
largeareafromfrontaltoparietal(excludingtemporalareas)fortheP3a
tasks(asreportedin[21]).Employingspatialmasksismoreappropriate
fortestingspecifichypothesesbasedonpreviousresearchasopposedto
exploratory data analysis. Use of spatial masks reduces the number of
statistical comparisons, and therefore the Type I error rate.
3. Results
The mean percentage of retained epochs of all the P3 ERP tasks at
Month 0 was 91.45% and 84.69% for the control and MS groups,
respectively.AtMonth12,themeanpercentageofretainedepochsofall
the P3 ERP tasks was 92.92% and 87.09% for the control and MS groups,
respectively.TheMSpatientsandcontrolsdidnotdifferinage(pN0.05).
EDSSscoredidnotchangesignificantlyover12-monthperiod(pN0.05).
HRSD-21 score had not changed in either group over 12-month period
but there was a group effect as MS patients had a higher score than the
controls at both time points (pb0.05). MS patients had lower level of
education than controls (pb0.05) and therefore the PASAT 12-month
difference score was controlled for education level in the SPM analyses.
A repeated measures ANOVA was conducted on PASAT and reaction
time data. There was a group effect for the PASAT score as MS patients
performed worse in both time points (F[1,37]=4.22, p=0.047). At a
group level, there was no significant difference in the PASAT score
betweenMonth0andMonth12ineitherMSpatientsorcontrols.RRMS
and SPMS patients did not differ in the PASAT 12-month difference
score, or in the PASAT performance in Month 0 or in Month 12 (pN.05).
PASAT score did not correlate with EDSS or HRSD-21 scores at either
Month 0 or Month 12 (pN0.05). There were no statistically significant
reaction time differences between the two groups for either visual or
auditory modalities (pN0.05), but there was a time effect in the visual
P3b as both the controls and MS patients had longer reaction times at
Month 12 (F[1,41]=6.12, p=0.018).
47
H. Kiiski et al. / Journal of the Neurological Sciences 305 (2011) 45–52

Page 4

3.1. Correlation between the 12-month change in PASAT score and the
12-month change in P3 ERP amplitudes
SPM8, a statistical method for examining differences in EEG activity,
was used to examine scalp areas and time periods in which the
amplitude of 12-month ERP difference was significantly correlated with
the 12-month PASAT difference score. The correlations (Pearson's r) are
reported between the voxel of maximum intensity and the 12-month
PASAT difference score. For all subjects, the two difference estimates
werecorrelatedinbothauditoryandvisualP3atasks(r=0.48,p=0.003
and r=0.56, pb0.001, respectively), but not in auditory and visual P3b
tasks(Figs.1and2).Thecorrelations wereapparentalthoughatagroup
level the PASAT score had not significantly changed over the 12-month
period.ThechangeinthePASATscorewasvariableinindividualsubjects
withsomesubjects'PASATscoreincreasing,someremainingsimilarand
some reducing (standard deviation 9.57% in MS and 9.36% in controls).
Therefore, the change in the PASAT score for individual subjects was
paralleled by the individual changes in P3a ERP amplitudes.
There was a significant correlation between the voxel of maximum
intensity and the PASAT score at Month 0 for both auditory and visual
P3aERPs(r=0.56,pb0.001andr=0.71,pb0.001,respectively),andat
Month12 for auditory and visual P3a ERP/PASAT correlation in MS
patients (r=0.63, pb0.05 and r=0.61, pb0.05, respectively). In
controls, the SPM for the correlation at Month12 displayed a trend
towardsstatisticalsignificance (pb.005). ThePASAT results forcontrols
atMonth12 were less variablethanMonth0,and the average scorewas
higher. Therefore, the results from the control group alone at Month 12
did not have enough variability for an accurate correlation.
3.2. Longitudinal comparison of P3 ERP amplitudes
Whenthe12-monthdifferenceimageswere comparedbetween MS
patients and controls, MS patients had greater reduction in amplitude
thanthecontrolsatalargecentro-parietalarea242–441 msinthevisual
P3b task, and over a more restricted parietal area 556–598 ms in the
visual P3a task (Fig. 3). Paired t-tests of the MS patients revealed
reduced amplitudes at Month 12 compared to Month 0 in visual P3b,
auditory P3b and auditory P3aover the centro-parietal area (Fig. 4). For
the controls, the amplitudes were observed to be reduced in auditory
P3a task at Month 12 over frontal area 243–286 ms and in visual P3a
task over central area at 266–306 ms and 368–446 ms (Supplemental
Fig. 1). Grand averaged waveforms of the responses to visual and
auditory P3b (Pz) and P3a (FCz) for Month 0 versus Month 12 in MS
patients and controls are displayed in Fig. 5.
4. Discussion
4.1. The results of this study
The present study employed a high-density EEG array across both
visual and auditory modalities with MS patients and controls over 12-
month period. In the present study the 12-month difference in PASAT
score correlated statistically significantly with the 12-month difference
in amplitude of visual P3a over the parietal area, which was consistent
topographically with a previous study [21]. A similar correlation was
observed with the 12-month difference in amplitude of auditory P3a.
The correlations were evident even though the PASAT score – at the
group level – had not significantly changed over the 12-month period.
The change in the PASAT score was variable across subjects, with some
subjects' PASAT score increasing, some remaining similar and some
reducing (standard deviation 9.57% in MS and 9.36% in controls). The
finding that cognitive function changes over time in MS may be
reflectiveofthediseasecoursevariability.Therefore,itisinterestingthat
P3 ERP amplitudes reflect changes in cognitive function measured by
the PASAT score.
The present study also revealed the 12-month difference amplitude
to be statistically significantly greater for MS patients compared to
healthy age-matched controls, and the effect was best demonstrated
with visual P3b task over a large centro-parietal region. Similarly, the
decrease in amplitudes over 12-month period was greater in MS
patients, relative to controls, over parietal area for visual P3a. The
amplitudes of the visual P3b, auditory P3b and auditory P3a were
reduced in MS patients over the centro-parietal area after 12 months.
Visual P3a amplitudes were not significantly different in MS patients
from Month 0 toMonth 12. This mightbe due tothe fact that visualP3a
is a more difficult task compared to the P3b ERP tasks as reported in
previousstudies[9].Therefore,theamplitudesatMonth0mightalready
be at such a low level that a further decrease wasdifficult to measure at
Month12.Moredifficultstandard/targetdiscriminationsarereflectedin
later P3 onset [9,10,36]. Greater standard/target discrimination diffi-
culty in the visual P3a task – in comparison to the visual P3b task –
resultedinlongerP3aonset.Thereforesignificantdifferenceswerelater
Fig. 1. Significant correlation (pb0.005) between the 12-month amplitude changes of P3
event-related potentials and the 12-month change in PASAT with both MS patients and
controls. The correlation (Pearson's r) between the voxel of maximum intensity and the
12-month PASAT difference score is also displayed.
48
H. Kiiski et al. / Journal of the Neurological Sciences 305 (2011) 45–52

Page 5

for visual P3a versus visual P3b. In controls, the amplitudes of P3a ERP
tasks at Month 12 were reduced compared to Month 0 over frontal and
central areas, thus emphasizing the need for control groups in
longitudinal studies.
4.2. Comparison to previous studies
It is essential to recognise CI as early as possible and to monitor its
course frequently. The evolution of CI in MS has not been entirely
explained so far but a general trend appears to be that cognitive
impairment in MS occurs more slowly and less consistently in MS than
in other degenerative conditions such as Alzheimer's disease [5]. The
present study is the first one, to our knowledge, to investigate the
relationship between PASAT and P3 ERPs longitudinally in MS. The 12-
month PASAT difference score was found to correlate best with both
visualandauditoryP3aERP12-monthamplitudedifference.Priorcross-
sectional studies [11,21] have found correlation between PASAT and P3
ERPs, but they included larger number of subjects (N=54 and N=74,
respectively). Nevertheless, the additional power of the within-subject
design in the present study made it possible to observe statistically
significant PASAT/P3 ERP correlations for the within-subject change
scores.
There were some differences between the modalities as there was
greater variability in auditory P3a amplitude. Although the visual P3a
variability was less than the auditory P3a, perhaps reflecting the
difficulty of the visual P3a task, there was a sufficient change between
Month 0 and Month 12 for a statistically significant correlation. To be
more specific, the difficulty between the modalities is a function of
perceptual distinctiveness of the standard/target discrimination, and
not related to modality per se. Thus, the distinctiveness of the stimulus
and the context (standard/target discrimination) contributes to P3a
amplitude for the most part, and the modality of the stimuli is less
influential [37]. Therefore in our study the visual P3a task was more
difficult due to the stimulus properties (i.e. the standard/target
discrimination was more difficult for visual compared to auditory
stimuli).Nonetheless,therewasachangeforsomesubjectsinvisualP3a
amplitude over the 12-month period (see Fig. 2), and this change was
sufficient for a statistically significant correlation.
Cognitive electrophysiological measures have several advantages
overneuropsychologicalmeasures asthey are not adverselyaffectedby
practice effects [7], anxiety, education, motor delay of speech and/or
hand movement. More importantly, many patients dislike performing
the PASAT which leads to loss of data, and which is a serious
disadvantage especially in clinical trials in which repeated testing
mightincreasethedrop-outrate[38].Thereforeitcanbesuggestedthat
theP3aERPs may have utility, in parallel withneuropsychological tests,
in monitoring the course of subtle changes in CI in MS.
Fig. 2. Significant correlation (Pearson's r) between the voxel of maximum intensity and the 12-month PASAT difference score with both MS patients and controls.
Fig. 3. MS patients have significantly (pb0.005) greater decrease in amplitude over
12-month period compared to controls in visual P3b and P3a.
49
H. Kiiski et al. / Journal of the Neurological Sciences 305 (2011) 45–52

Page 6

Differences in P3 ERP amplitudes were observed over the 12-month
periodinthepresentstudy.OnlyonepreviouslongitudinalP3ERPstudy
[24]reported change in P3b component after one year whereas therest
ofthestudies[19,25,26]didnot.Theresultsofpreviousstudiesmightbe
confounded by the variety of ways that latency and amplitude were
measured. A high-density EEG array covers the entire scalp area and
therefore it is possibleto generate statistical parametric maps depicting
activations across time and the whole scalp area. Prior studies have
suggested a modality-specific amplitude reduction in MS [39]. In the
presentstudy,visualmodalityappearedtobemoresensitivetochanges
in the amplitudes of P3 ERPs as greater decrease was evident in MS
patients, relative to controls, over 12-month period. Similar distinction
between the modalities have been reported earlier [21,39], which may
be due to the differences between visual and auditory processing
pathways. There may be more demyelination in visual processing areas
and/or visual tracts, which are longer than auditory tracts. The
superiority of visual condition in detecting CI-related changes in P3
ERPscouldexplaininpartwhymostofthepreviouslongitudinalP3ERP
studies (except for [25]), which employed only auditory P3 paradigm,
failed to find any changes. The topographical differences in the
longitudinal changes of P3 ERP amplitudes require further study;
especially the changes in the P3 ERPs over a longer period of time and
the differences between modalities would need more investigation.
Our study is the first longitudinal P3 ERP study to our knowledge to
focus solely on the changes in P3 ERPs related to CI in MS, and not
concentrate on the effect of medication on cognition. The present study
comprised of a more comprehensive patient group – including SPMS
patients – who have not been part of the previously reported
longitudinal P3 ERP studies (e.g. [24]). The results of the present study
might be influenced by the different medications the patients were
receiving. The disease-modifying medications alter the cerebral demy-
elinatingprocess,whichmightinturnbeassociatedwithslowerdecline
of cognitive functions [40]. On the other hand, this link has not been
establishedfirmly,andthemixofMSpatientswithdifferenttreatments
has the advantage of giving a more comprehensive overview of the
process of CI in MS. In the present study it was not possible to compare
the patients receiving disease-modifying medication to those without
medication as the number of patients without medication for MS was
Fig. 4. MS patients have significantly (pb0.005) decreased amplitudes at Month 12 compared to Month 0 in visual P3b, auditory P3b and auditory P3a.
50
H. Kiiski et al. / Journal of the Neurological Sciences 305 (2011) 45–52

Page 7

too low (N=6). Descriptive statistics of PASAT 12-month difference
scores for these two groups are presented in Supplemental Table 1.
Increasing brain lesion load and brain atrophy are correlated with
the progression of CI in MS [5]. Furthermore, the whole brain atrophy
seems to be a good predictor of future cognitive and memory decline
in early MS [41]. The relationship between lesions, and neuropsycho-
logical and neurophysiological cognitive function is imprecisely
defined in MS [42], which is especially true concerning progression
of the CI in MS. Future ERP studies could also examine abnormalities
in normal-appearing white and grey matter, which may be related to
the severity and progression of CI but are commonly undetectable by
conventional imaging [43]. Incorporating a full neuropsychological
battery to longitudinal ERP and MRI studies could clarify how the
changes in specific cognitive domains are related to longitudinal
neurophysiological and/or MRI findings.
Supplementary materials related to this article can be found online
at doi:10.1016/j.jns.2011.03.018.
Acknowledgements
This study was funded by an Enterprise Ireland (eBiomed:
eHealthCare based on Biomedical Signal Processing and ICT for
Integrated Diagnosis and Treatment of Disease) and Science Founda-
tion Ireland (09/RFP/NE2382) grants to R.B. Reilly, and an IRCSET
EMBARK Postgraduate Scholarship to H. Kiiski.
References
[1] Hoffmann S, Tittgemeyer M, von Cramon DY. Cognitive impairment in multiple
sclerosis. Curr Opin Neurol 2007 Jun;20(3):275–80.
[2] Rogers JM, Panegyres PK. Cognitive impairment in multiple sclerosis: evidence-
based analysis and recommendations. J Clin Neurosci 2007 Oct;14(10):919–27.
[3] Rao SM, Leo GJ, Ellington L, Nauertz T, Bernardin L, Unverzagt F. Cognitive
dysfunction in multiple-sclerosis. 2. Impact on employment and social function-
ing. Neurology 1991 May;41(5):692–6.
[4] Kalmar JH, Halper J, Gaudino EA, Moore NB, DeLuca J. The relationship between
cognitive deficits and everyday functional activities in multiple sclerosis.
Neuropsychology 2008 Jul;22(4):442–9.
[5] Amato MP, Zipoli V, Portaccio E. Multiple sclerosis-related cognitive changes: a
reviewofcross-sectionalandlongitudinalstudies.JNeurolSci2006Jun15;245(1–2):
41–6.
[6] Cutter GR, Baier ML, Rudick RA, Cookfair DL, Fischer JS, Petkau J, et al. Development
of a multiple sclerosis functional composite as a clinical trial outcome measure.
Brain 1999 May;122(Pt 5):871–82.
[7] Barker-Collo SL. Within session practice effects on the PASAT in clients with
multiple sclerosis. Arch Clin Neuropsych 2005 Mar;20(2):145–52.
[8] Leocani L, Locatelli T, Martinelli V, Rovaris M, Falautano M, Filippi M, et al.
Electroencephalographic coherence analysis in multiple sclerosis: correlation with
clinical, neuropsychological, and MRI findings. J Neurol Neurosurg Psychiatry 2000
Aug;69(2):192–8.
[9] Polich J. Updating P300: an integrative theory of P3a and P3b. Clin Neurophysiol
2007 Oct;118(10):2128–48.
[10] Kok A. On the utility of P3 amplitude as a measure of processing capacity.
Psychophysiology. 2001 14 March 2003;38(3):557–77.
[11] Whelan R, Lonergan R, Kiiski H, Nolan H, Kinsella K, Bramham J, et al. A high-
density ERP study reveals latency, amplitude, and topographical differences in
multiple sclerosis patients versus controls. Clin Neurophysiol 2010 Sep;121(9):
1420–6.
[12] KiiskiH, Whelan R, Lonergan R,Nolan H, KinsellaK, Hutchinson M, et al. Preliminary
evidenceforcorrelationbetweenPASATperformanceandP3aandP3bamplitudesin
progressive multiple sclerosis. Eur J Neurol 2010 [Short communication]; doi:
10.1111/j.1468-1331.2010.03172.x.
[13] Aminoff JC, Goodin DS. Long-latency cerebral event-related potentials in multiple
sclerosis. J Clin Neurophysiol 2001 Jul;18(4):372–7.
[14] Ellger T, Bethke F, Frese A, Luettmann RJ, Buchheister A, Ringelstein EB, et al.
Event-related potentials in different subtypes of multiple sclerosis—a cross-
sectional study. J Neurol Sci 2002 Dec 15;205(1):35–40.
[15] Gil R, Zai L, Neau JP, Jonveaux T, Agbo C, Rosolacci T, et al. Event-related auditory
evoked potentials and multiple sclerosis. Electroencephalogr Clin Neurophysiol
1993 May–Jun;88(3):182–7.
[16] Piras MR, Magnano I, Canu ED, Paulus KS, Satta WM, Soddu A, et al. Longitudinal
study of cognitive dysfunction in multiple sclerosis: neuropsychological, neuro-
radiological, and neurophysiological findings. J Neurol Neurosurg Psychiatry 2003
Jul;74(7):878–85.
[17] Polich J, Romine JS, Sipe JC, Aung M, Dalessio DJ. P300 in multiple sclerosis: a
preliminary report. Int J Psychophysiol 1992 Mar;12(2):155–63.
[18] Sailer M, Heinze HJ, Tendolkar I, Decker U, Kreye O, v Rolbicki U, et al. Influence of
cerebral lesion volume and lesion distribution on event-related brain potentials in
multiple sclerosis. J Neurol 2001 Dec;248(12):1049–55.
[19] Gerschlager W, Beisteiner R, Deecke L, Dirnberger G, Endl W, Kollegger H, et al.
Electrophysiological, neuropsychological and clinical findings in multiple sclerosis
patients receiving interferon beta-1b: a 1-year follow-up. Eur Neurol 2000;44(4):
205–9.
[20] Jung J, Morlet D, Mercier B, Confavreux C, Fischer C. Mismatch negativity (MMN)
in multiple sclerosis: an event-related potentials study in 46 patients. Clin
Neurophysiol 2006 Jan;117(1):85–93.
[21] Whelan R, Lonergan R, Kiiski H, Nolan H, Kinsella K, Hutchinson M, et al. Impaired
information processing speed and attention allocation in multiple sclerosis
patients versus controls: a high-density EEG study. J Neurol Sci 2010 Jun 15;293
(1-2):45–50.
[22] Leocani L, Comi G. Neurophysiological markers. Neurol Sci 2008 Sep;29:S218–21.
[23] Pelayo R, Montalban X, Minoves T, Moncho D, Rio J, Nos C, et al. Do multimodal
evoked potentials add information to MRI in clinically isolated syndromes? Mult
Scler 2010 Jan;16(1):55–61.
[24] Flechter S, Vardi J, Finkelstein Y, Pollak L. Cognitive dysfunction evaluation in
multiple sclerosis patients treated with interferon beta-1b: an open-label
prospective 1 year study. Isr Med Assoc J 2007 Jun;9(6):457–9.
Fig. 5. Grand averaged waveforms of the responses to visual and auditory P3b (Pz) and P3a (FCz) for Month 0 versus Month 12 in MS patients and controls.
51
H. Kiiski et al. / Journal of the Neurological Sciences 305 (2011) 45–52

Page 8

[25] Pliskin NH, Hamer DP, Goldstein DS, Towle VL, Reder AT, Noronha A, et al.
Improved delayed visual reproduction test performance in multiple sclerosis
patients receiving interferon beta-1b. Neurology 1996 Dec;47(6):1463–8.
[26] Pokryszko-Dragan A, Zagrajek M, Slotwinski K, Gruszka E, Bilinska M, Podemski R.
Neuropsychological testing and event-related potentials in the assessment of
cognitive performance in the patients with multiple sclerosis—a pilot study. Clin
Neurol Neurosur 2009 Jul;111(6):503–6.
[27] Polich J. Clinical application of the P300 event-related brain potential. Phys Med
Rehabil Clin N Am 2004 Feb;15(1):133–61.
[28] Polman CH, Reingold SC, Edan G, Filippi M, Hartung HP, Kappos L, et al. Diagnostic
criteria for multiple sclerosis: 2005 revisions to the “McDonald Criteria”. Ann
Neurol 2005 Dec;58(6):840–6.
[29] Tombaugh TN. A comprehensive review of the Paced Auditory Serial Addition Test
(PASAT). Arch Clin Neuropsychol 2006 Jan;21(1):53–76.
[30] Kurtzke JF. Historical and clinical perspectives of the expanded disability status
scale. Neuroepidemiology 2008;31(1):1–9.
[31] Vieweg BW, Hedlund JL. The Hamilton rating scale for depression. J Oper
Psychiatry 1979;10(2):149–65.
[32] Oostenveld R, Praamstra P. The five percent electrode system for high-resolution
EEG and ERP measurements. Clin Neurophysiol 2001 Apr;112(4):713–9.
[33] Whelan R. Effective analysis of reaction time data. Psychol Rec 2008;58(3):
475–82.
[34] Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial
EEG dynamics including independent component analysis. J Neurosci Meth 2004
Mar 15;134(1):9–21.
[35] Delorme A, Sejnowski T, Makeig S. Enhanced detection of artifacts in EEG data
using higher-order statistics and independent component analysis. Neuroimage
2007 Feb 15;34(4):1443–9.
[36] Comerchero MD, Polich J. P3a, perceptual distinctiveness, and stimulus modality.
Cogn Brain Res 1998 Jul;7(1):41–8.
[37] Polich J. Task-difficulty, probability, and inter-stimulus interval as determinants of
P300 from auditory-stimuli. Electroen Clin Neuro 1987 Jul;68(4):311–20.
[38] Williams J, O'Rourke K, Hutchinson M, Tubridy N. The face-symbol test and the
symbol-digit test are not reliable surrogates for the Paced Auditory Serial Addition
Test in multiple sclerosis. Mult Scler 2006 Oct;12(5):599–604.
[39] De Sonneville LM, Boringa JB, Reuling IE, Lazeron RH, Ader HJ, Polman CH.
Information processing characteristics in subtypes of multiple sclerosis. Neurop-
sychologia 2002;40(11):1751–65.
[40] Bobholz JA, Rao SM. Cognitive dysfunction in multiple sclerosis: a review of recent
developments. Curr Opin Neurol 2003;16(3):283–8 [review].
[41] Whitwell JL. Longitudinal imaging: change and causality. Curr Opin Neurol 2008
Aug;21(4):410–6.
[42] Lazeron RHC, Boringa JB, Schouten M, Uitdehaag BMJ, Bergers E, Lindeboom J, et al.
Brain atrophy and lesion load as explaining parameters for cognitive impairment
in multiple sclerosis. Mult Scler 2005 Oct;11(5):524–31.
[43] Bagert B, Camplair P, Bourdette D. Cognitive dysfunction in multiple sclerosis—
natural history, pathophysiology and management. CNS Drugs 2002;16(7):445–55.
52
H. Kiiski et al. / Journal of the Neurological Sciences 305 (2011) 45–52