Anda di halaman 1dari 10

J Neurol (2012) 259:13371346 DOI 10.



Increased cerebral activation after behavioral treatment for memory decits in MS

Nancy D. Chiaravalloti Glenn Wylie Victoria Leavitt John DeLuca

Received: 25 May 2011 / Revised: 30 November 2011 / Accepted: 1 December 2011 / Published online: 12 January 2012 Springer-Verlag 2012

Abstract Decits in new learning and memory are common in persons with multiple sclerosis (MS), though few studies have examined the efcacy of memory retraining in MS. Previous research from our laboratory has demonstrated that the modied Story Memory Technique (mSMT) signicantly improves new learning and memory in MS. The present double-blind, placebo-controlled, randomized clinical trial was designed to examine changes in cerebral activation following mSMT treatment. Sixteen individuals with clinically denite MS were randomly assigned to treatment (n = 8) or placebo-control (n = 8) groups, matched for age, education, and disease characteristics. Baseline and follow-up fMRI was collected during performance of learning and memory tasks. No baseline activation differences on fMRI were seen between groups. After treatment, greater activation was evident in the treatment group during performance of a memory task within a widespread cortical network involving frontal, parietal, precuneus, and parahippocampal regions. All

participants in the treatment group showed increased activation in frontal and temporal regions in particular. In contrast, the control group showed no signicant changes in cerebral activation at follow-up. A signicant association was found between increased activation in the right middle frontal gyrus and improved memory performance post-treatment. The increased activation seen likely reects increased use of strategies taught during treatment when learning new information. This study is the rst to demonstrate a signicant change in cerebral activation resulting from a behavioral memory intervention in an MS sample. Behavioral interventions can show signicant changes in the brain, validating clinical utility. Keywords Cognition fMRI Multiple sclerosis Memory Cognitive rehabilitation

Introduction Multiple sclerosis (MS) is a progressive disease of the central nervous system (CNS) characterized by widespread lesions in the myelin sheath of the CNS and damage to gray matter [1]. A wide symptom array follows, including cognitive impairment [2]. Learning and memory is the most common cognitive symptom [2, 3], evident in 4065% of patients [4] and demonstrated to be a result of new learning decits [5, 6]. The treatment of memory decits has received little attention in MS [7]. Imagery is a widely utilized cognitive intervention for memory problems with much data supporting its efcacy [7, 8]. Our randomized clinical trial examining the efcacy of the modied Story Memory Technique (mSMT) in MS studied 29 MS participants with new learning decits who were randomly assigned to an

N. D. Chiaravalloti G. Wylie V. Leavitt J. DeLuca Kessler Foundation Research Center, West Orange, NJ, USA N. D. Chiaravalloti G. Wylie V. Leavitt J. DeLuca Department of Physical Medicine and Rehabilitation, UMDNJ, New Jersey Medical School, Newark, NJ, USA N. D. Chiaravalloti (&) Neuropsychology and Neuroscience Laboratory, Kessler Foundation Research Center, 300 Executive Drive, Suite 70, West Orange, NJ 07052, USA e-mail: J. DeLuca Department of Neurology and Neurosciences, UMDNJ, New Jersey Medical School, Newark, NJ, USA



J Neurol (2012) 259:13371346

experimental or placebo control group [9]. Results indicated that the mSMT signicantly improved new learning, particularly in those with a baseline moderate/severe new learning impairment. Functional neuroimaging in MS Studies examining working memory in MS using fMRI have shown bilateral frontal cortical activation, as well as more dispersed activation throughout the brain [1012]. This is in contrast to the more focal left frontal activation seen in controls. Distinct patterns of cerebral activation are found in MS samples with cognitive impairment relative to those without [10, 11]. This indicates that patterns of cerebral activation are altered in MS on cognitive tasks and this alteration is specic to those demonstrating impairment. Thus, improvements in cognition, independent of the disease process, such as in response to cognitive rehabilitation, will likely produce changes in patterns of cerebral activation [13]. Applying neuroimaging to cognitive rehabilitation Initial intervention studies suggest that in response to training, the brain shows both increases and decreases in cerebral activity [14] and one can observe changes in brain functions with a brief behavioral intervention. While no studies to our knowledge have examined changes in brain activation in cognition following cognitive rehabilitation in MS, studies have examined changes post-intervention in the motor domain in MS [15, 16]. In controls, Poldrack et al. [17] showed that practiced mirror reading resulted in decreased activation in occipital, inferior temporal, and superior parietal cortices and increased activation in occipito-parietal and lateral temporal regions on fMRI. Similarly, Olesen et al. [18] showed both increases (e.g., parietal and frontal cortices) and decreases (e.g., anterior cingulate) in brain regions after a 5-week training with a working memory paradigm in controls. Such research is yet to be completed in neurological samples. The present study is a double-blind, placebo-controlled, randomized clinical trial to examine changes in patterns of cerebral activation on new learning and memory tasks administered during fMRI from before to after treatment with the mSMT in persons with MS. To our knowledge, only one study to date has examined changes on fMRI following cognitive rehabilitation in a neurological sample. Belleville et al. [19] examined the effect of memory retraining on patterns of brain activation on fMRI in a sample with mild cognitive impairment (MCI), in contrast to a group of healthy controls undergoing the same treatment. While the HC group showed decreased activation after training, the MCI group showed increased activation

following training within a large network including the frontal, temporal, and parietal regions. The current study examines a similar question in a sample of persons with MS. That is, does a cognitive intervention targeting learning and memory decits with proven behavioral efcacy also elicit change on the neural level in memoryimpaired persons with MS? We hypothesized that subjects in the experimental group would show a signicant change in patterns of cerebral activation when performing an fMRI task assessing new learning abilities via a list-learning paradigm. Specically, we expected that the experimental group would show increased activation in (1) posterior regions of the right hemisphere as well as (2) bilateral regions of the frontal lobe at immediate follow-up relative to baseline. These changes were expected due to the engagement of imagery procedures to facilitate learning, as well as the increased use of context to organize incoming information. No changes in patterns of cerebral activation were expected to be observed in the placebo-control group.

Method Participants Participants were 16 individuals with clinically denite multiple sclerosis (MS), according to the criteria of McDonald [20]. Eight individuals were randomly assigned to the treatment group and eight individuals were randomly assigned to the placebo-control group through the use of a computerized random number generator. There were no signicant differences between the groups in terms of age, education, estimated premorbid IQ, pre-treatment new learning ability, or gender. The groups were similar in terms of disease characteristics including disease duration (in months), months since symptom onset, months since most recent exacerbation, disease subtype, and ambulation index (Table 1). Methods All potential participants rst signed an informed consent form approved by the Institutional Review Board of Kessler Foundation Research Center and UMDNJ. Prior to enrollment in the study protocol, all potential participants underwent a two-part screening: (1) an initial screening examination via telephone that screened for age, most recent exacerbation, neurological history, medications, and MRI compatibility. Patients who passed the initial screening then completed a (2) detailed, in-person screening that evaluated psychiatric and substance-abuse history, visual acuity, language comprehension and new learning and memory abilities. Persons with a history of major depressive disorder, schizophrenia, bipolar disorder I or II


J Neurol (2012) 259:13371346 Table 1 Demographic and pre-treatment variable by group (all comparisons non-signicant) Treatment group M (SD) n = 8 Age Education WASI vocabulary (estimate of premorbid IQ) Gender Pretreatment learning ability (OT-SRT trials to criterion) Disease duration (in months) Months since symptom onset Months since most recent exacerbation Ambulation index Disease subtype 49.25 (9.33) 16.38 (3.20) 8.67 (4.55) 7/8 female 10.75 (2.43) 186.71 (116.95) 188.4 (168.3) 30 (24.4) 2.13 (1.73) 5/8 RR Control group M (SD) n = 8 46.75 (6.27) 16.5 (1.77) 11.88 (2.59) 7/8 female 11.0 (3.02) 177.14 (66.49) 202.8 (56.45) 11.5 (2.12) 3.75 (1.39) 6/8 RR


Value t = -0.63 t = 0.097 t = 1.68 X2 = 0 t = 0.18 t = -0.19 t = 0.19 t = 1.01 t = 2.07 X2 = 0.26

were excluded from the study. All participants had to demonstrate sufcient visual acuity to see the test materials. Therefore, persons whose vision was signicantly impaired by scotomas (corrected vision in worse eye \20/60), diplopia, or nystagmus were excluded. In addition, all participants were right handed to control for the inuence of handedness on brain organization. All participants were required to demonstrate intact language comprehension, as documented by a Token Test score above 26, demonstrating an intact ability to follow one-, two-, and three-step commands. In addition, to qualify for study participation, participants were required to demonstrate new learning and memory abilities that were at least 1.5 standard deviations below the mean of a healthy control group, as documented by the Open-Trial Selective Reminding Test (OT-SRT; [21]). That is, only individuals that required eight or more trials to reach the learning criterion on the open-trial SRT were included in the study, ensuring that all subjects in the study had memory impairments.
Fig. 1 Experimental overview

Upon satisfaction of the inclusion criteria, participants were randomly assigned to the treatment group or a placebo-control group. All participants completed baseline testing, consisting of a neuropsychological assessment, questionnaires to assess everyday life cognitive abilities, and functional magnetic resonance imaging (fMRI) prior to beginning treatment, as well as the same procedures with alternate forms of memory tests within 1 week of completing treatment. The same research assistant conducted both baseline and immediate follow-up evaluations, referred to as raters; these assistants were blind to group membership as they did not provide treatment in this study (see Fig. 1 for treatment ow chart). The intervention for both the treatment and placebo groups was provided by research assistants, other than those who conducted the preand post-assessments, referred to as treaters. It was assured that raters were blind to group membership via several mechanisms: (1) Raters and treaters were different; (2) Each rater and treater had his/her own database to enter data on treatment or assessments. This data was assembled



J Neurol (2012) 259:13371346

into its full form by the research manager, who did not conduct either treatment or assessment sessions. The raters NEVER saw the treatment data and the treaters NEVER saw the assessment data; (3) Treaters and raters were strictly instructed NOT to discuss their participants. All questions/communications were directed through the research manager, who once again did not see patients for interventions or assessments; (4) Participants were instructed NOT to discuss any aspect of their treatment with the person conducting their assessment. Both treatment and assessment sessions were randomly observed through a one-way mirror for two purposes: (1) to assure treatment integrity, and (2) to monitor the interaction between participant and rater to assure blinding was maintained. Raters were then questioned by the research manager as to any information provided by the participant regarding treatment activities. In no case did a participant discuss treatment activities with their raters, as they were instructed not to do this. The integrity of blinding procedures was tested via a questionnaire administered after completion of the longterm follow-up for all participants. Participants were specically asked which group they thought they were in (treatment or control). Data were available for nine of the 16 participants in the current study. Of these nine participants, ve (55%) said that they were in the group that they were indeed assigned to (3/4 controls stated that they were in the control group and 2/5 treatment participants stated that they were in the treatment group). This data indicates that the blinding procedures were effective such that when participants were forced to guess which group they were in, the correct guess rate fell at chance levels. The study was submitted to prior to enrolling the rst participant. Neuroimaging procedures

was realigned to the rst remaining image of the rst series. The images were smoothed using an 8-mm3 Gaussian smoothing kernel, and scaled to the mean intensity. The data were then deconvolved using a boxcar function in which each condition was represented by a regressor (motion parameters and three polynomial regressors [to model signal drift] were included as regressors of no interest). The results of the deconvolution were then warped into standard space. This was done by rst warping the high-resolution MPRAGE image into standard space and then applying the same non-linear warping parameters to the results of the deconvolution (i.e., the beta weights for each condition). This was done for each subjects data individually. Once the data from all subjects were in standard space, group-level statistics (ANOVAs) were performed, with factors of time (pre- and post- treatment) and group (treatment and control). The threshold for signicance was set at an alpha of 0.01 and multiple comparisons were controlled for by requiring a cluster size of 10 voxels in native space (3.438 9 3.438 9 4 mm), or 473 voxels after the data were resampled to 1-mm isotropic voxels. This technique for setting the cluster threshold controls false positives while also sparing statistical power [23]. As a follow-up to the activation analysis, correlational analyses were conducted for the treatment group to examine the association between fMRI activation posttreatment and memory performance post-treatment, controlling for activation and memory pre-treatment. In the regions of frontal and temporal cortex that showed increased activation in the treatment group, partial correlations were conducted between the post-intervention beta weights (fMRI data) and the post-intervention memory performance (behavioral data), controlling for baseline activation and baseline memory performance. fMRI tasks

fMRI data was acquired on a Siemens Allegra 3T MR scanner. Images were acquired using a standard RF head coil. Contiguous slices (3.438 mm2 in-plane resolution; slice thickness = 4 mm) were obtained parallel to the ACPC line. Structural images were also acquired, using a standard T1-weighted pulse sequence: an MPRAGE (echo time = 4.38 ms; repetition time = 2,000 ms, eld of view = 220 mm; ip angle = 8; slice thickness = 1 mm, number of excitations = 1, matrix = 256 9 256, in-plane resolution = 0.859 mm2). Functional images were acquired using gradient-echo echo-planar imaging (echo time = 30 ms; repetition time = 2,000 ms; eld of view = 22 cm; ip angle = 80). For each time series, the rst ve images were discarded to ensure steady-state magnetization. All images were preprocessed using AFNI [22]. Each time-series of images

Word learning task A list-learning task was completed within the scanner, administered at both testing sessions (pre-treatment, posttreatment), utilizing an alternate form for the post-treatment session. This task utilizes a blocked design in which participants were presented with monochromatic printed words using standard fMRI-compatible Avotec Head Coil Montage designed for visual stimulus delivery. Stimulus delivery was regulated using E-prime software and a PCcompatible laptop computer. Prior to each stimulus presentation, a xation cross was presented for 500 ms in order to prompt the participant to attend to the visual display. A stimulus (i.e., a word) was then presented for 2 s, followed by a 2-s inter-stimulus interval (ISI). Participants


J Neurol (2012) 259:13371346


were required to read the word (subvocally), and to press the response button during the ISI if they had a pleasant association to the word. This approach has been taken in other fMRI studies [24] as a means of promoting deeper encoding of the stimuli. Sixteen words were presented in each encoding block, for a total block time of 32 s. The four encoding blocks were separated by blocks during which the stimulation parameters were identical, but during which XXXXX was presented instead of the words to be encoded. Subjects were asked to press the response button when the Xs were presented. Word recognition task Participants were again presented with monochromatic printed words using the stimulus delivery equipment and software described above. Prior to each stimulus presentation, a xation cross was presented for 500 ms in order to prompt the participant to attend to the visual display. A stimulus (i.e., a word) was then presented for 2 s, followed by a 2-s ISI. Participants were required to read the word (subvocally), and to press a button during the ISI as follows: (1) the left response button if they recalled the word as having been previously presented during the encoding task or (2) the right response button if the word was not presented during the encoding task. Of the 16 words from the original list, each participant was presented with eight of the target words from the original 16 interspersed with eight novel words that served as foils. This was done in order to keep the length of the recall session consistent with the length of encoding. The four recognition blocks were separated by blocks with identical stimulation parameters, but during which XXXXX was presented instead of words to be recognized. As in the encoding blocks, subjects were instructed to press the button when the Xs were presented. Treatment protocol The treatment protocol consisted of ten sessions of the modied Story Memory Technique (mSMT), which has been previously shown to be effective for treating new learning and memory decits in an MS population [9]. Treatment consisted of the SMT training twice per week for 5 weeks, lasting 4560 min each. The SMT entails two related skills: (1) imagery and (2) context. During sessions 14, subjects were taught to utilize imagery to facilitate the learning of verbal information. Sessions 58 taught participants to utilize context to facilitate learning. Given the fact that real life rarely requires that one remember a list of words, sessions 9 and 10 focus on applying the mSMT to real-world settings. Real-life situations were addressed. If the participant could not describe two memory-taxing

situations unique to their life, real-life situations were provided for them, from which they chose two situations they would most likely encounter, including (1) remembering a lengthy shopping list, (2) recalling a list of errands, and (3) recalling steps in driving directions. The treatment is highly manualized and the therapist therefore followed a training manual with scripts provided. The control group met with the treaters at the same frequency as the experimental group, but engaged in verbal tasks to control for professional contact and alterations in the disease process. Tasks consisted of reading the same stories that the experimental group used and answering questions. The placebo task was matched to the training task for duration of contact with the treater, and medium of presentation, specically via computer. The only difference between the two groups is that the mSMT training was only provided to subjects in the experimental group.

Results Behavioral results On the CVLT short-delay free recall, there was a signicant difference between the treatment and control groups in terms of their behavioral response to the intervention. A greater number of individuals in the treatment group showed greater than a 10% improvement on CVLT shortdelay free recall from pre- to post-treatment than the control group (X2 (1) = 4.0, p \ 0.05; Table 2). We chose 10% as an indicator of improvement due to the fact that the existing literature on pharmacotherapy for improving memory functioning in MS generally documents a 10% improvement in the treatment group (see [25, 26]). Imaging results The fMRI data were rst analyzed with a mixed, betweenand within- subjects 2 9 2 ANOVA. The between-subjects factor was the group (treatment vs. control) and the withinsubjects factor was time (pre- vs. post-treatment for the treatment group, pre- vs. post-sham treatment for the controls). Several planned contrasts as part of this ANOVA were also performed. The rst compared cerebral activity in the two groups prior to any intervention (treatment vs.

Table 2 Behavioral changes post-intervention in the treatment group versus the control group (X2 (1) = 4.0, p \ 0.05) Improvement Treatment Control 6 2 No improvement 2 6



J Neurol (2012) 259:13371346

control group at baseline); the second compared activity before and after treatment in the control group; the last compared activity before and after treatment in the treatment group. Baseline scans revealed no signicant differences in cerebral activation between the treatment and control groups during completion of the word learning task. Examination of the control group activation at post-treatment revealed no differences relative to pre-treatment activation maps. In contrast, the fMRI data collected for the treatment group at post-treatment compared to pre-treatment revealed signicantly increased activation in multiple cortical areas (Table 3; Fig. 2). Importantly, an examination of individual subject data for the treatment group revealed that increased activation subsequent to the intervention was evident in every subject (8 of 8) within the frontal and temporal cortices, areas related to learning and memory. Of primary interest in this study was any region showing an interaction between group and time such that activity changed for the treatment group but not for the control group. This pattern of results (i.e., an interaction) was found in a distributed network of areas including areas associated with memory (hippocampal and temporal areas), those associated with executive control (frontal and parietal areas), those associated with visual processing (occipital areas), as well as cerebellar regions. These areas are reported in Tables 3 and 4 (see also Fig. 2). Imaging-behavior association results Within the treatment group only, partial correlations were run to examine the relationship between differences in activation from pre- to post-treatment and changes in memory performance. That is, we examined the
Table 3 Treatment group differences in fMRI activation at post-treatment

relationship between post-intervention activation in each of the ve regions showing increased activation in frontal and temporal cortices, and performance at post-intervention on the CVLT short-delay free recall. Controlling for baseline activation and baseline performance, a positive association was found for right middle frontal gyrus and total responses provided on the CVLT short-delay free recall (r = 0.73, p = 0.05; all tests were one-tailed). That is, increased activation in right MFG was accompanied by improved performance in the treatment group.

Discussion The present double-blind, placebo-controlled, randomized clinical trial provides the rst evidence of changes at the neural level accompanying behavioral improvements following completion of a behavioral intervention targeting learning and memory in persons with MS with cognitive impairment. Prior to this treatment, the two groups of MS participants were comparable both in terms of neuropsychological performance and cerebral activity. However, after treatment, changes on the neural level were evident in the treatment group relative to controls. Specically, areas of increased activation were seen in frontal, temporal, parietal, occipital, and cerebellar regions at post-treatment compared with pre-treatment. In contrast, no activation increases from baseline to follow-up were noted in the placebo-control group. Moreover, while our group level results indicated increased activation in areas related to memory and learning, importantly, every one of the treatment group participants showed a pattern of increased activation in areas of frontal and temporal cortices subsequent to the intervention. These increases in activation in areas of the brain known to be involved in memory,

Location Frontal Middle frontal gyrus Middle frontal gyrus Inferior frontal gyrus Parietal Precuneus Precuneus Inferior parietal lobule Posterior cingulate



T score

9 9 47 7 7 40 29 20 21

-30 26 30 10 -26 42 -2 54 -54 18

23 31 27 -73 -69 -49 -41 -37 -41 -53

28 28 -4 40 52 52 16 -8 -4 -36

51 14 12 26 16 20 24 15 14 18

3.98 4.84 4.35 5.08 3.77 3.66 3.62 3.71 4.04 4.56

All differences were positive, indicating increased activation after treatment in areas shown (p \ 0.01). An examination of control group at pre- versus post-treatment revealed no signicant differences

Temporal Middle temporal gyrus Middle temporal gyrus Cerebellar Cerebellar tonsil


J Neurol (2012) 259:13371346


Fig. 2 Results of the 2 9 2 ANOVA with factors of time and group. Following treatment, signicant increases in activation were seen in the treatment group relative to the control group in regions including frontal lobe, parietal lobe, and cerebellum. All comparisons are signicant at p \ 0.01 (minimum cluster size = 10 voxels). a Bold activation change from pre- to post-treatment in parahippocampal

gyrus. Control group represented by blue line; treatment group represented by red line. All interactions shown are signicant at p \ 0.01. b Bold activation change from pre- to post-treatment in superior temporal gyrus. c Bold activation change from pre- to posttreatment in middle frontal gyrus. d Bold activation change from preto post-treatment in precuneus

imagery, and context were accompanied by a behavioral improvement (C10% increase) on a list-learning task from pre- to post-treatment in six of eight participants in the treatment group. This 75% rate of improvement in our treatment group is comparable to a 65% response rate in the treatment group in a trial of Donepezil to improve

memory functioning in MS [26]. Increased activation in multiple cortical regions shown in the present study is similar to the ndings of Belleville et al. [19] who similarly showed increased fMRI activation in a widespread network of areas in an MCI sample who underwent a similar cognitive intervention.


1344 Table 4 Signicant time 9 group interaction

J Neurol (2012) 259:13371346

Frontal Middle frontal gyrus Precentral gyrus Inferior frontal gyrus Parietal Superior parietal lobule Precuneus Precuneus Posterior cingulate Supramarginal gyrus Temporal Hypothalamus Superior temporal gyrus Superior temporal gyrus Cingulate gyrus Parahippocampal gyrus Cerebellar Cerebellar tonsil 19 -26 2 2 -46 -45 -33 -73 -81 -28 12 -28 -4 162 11 25 31 8.93 8.94 10.48 12.43 Culmen Pyramis of vermis Occipital Inferior occipital gyrus 25 22 22 23 30 6 -58 50 2 10 -1 -9 3 -29 -41 -8 0 0 28 -4 27 22 20 17 13 12.12 9.96 10.07 12.24 10.48 8 6 9 7 19 7 29 40 -22 42 -42 34 -34 14 2 -38 31 -1 3 -49 -65 -73 -41 -41 36 24 28 60 40 48 16 36 42 11 10 48 30 12 24 10 12.08 11.61 9.37 14.85 9.73 10.38 10.15 11.46

All signicances are explained by increased activation in the treatment group relative to the control group at post-treatment relative to pre-treatment

Changes in activation were seen in brain regions known to be associated with the skills taught in the current treatment paradigm, the mSMT. Specically, the mSMT teaches individuals to use context and imagery to facilitate learning. Regions of the frontal lobe showing increased activation at post-treatment included middle frontal gyrus, precentral gyrus, and inferior frontal gyrus. The frontal lobe is known to be involved in the organization of information into a more usable form (see [27]). During the course of treatment, participants learned to apply context to facilitate learning, which is an organizational strategy presumably relying on frontal lobe activity. Thus, it was expected that the frontal regions shown to be more active following treatment would be engaged to facilitate effective task completion. The increased activation seen thus likely reects greater recruitment of neural resources that were previously underused for performance of memory tasks. Given that the increase in activation of the right MFG was accompanied by and positively associated with an improvement on a memory task, the engagement of these areas subsequent to training likely indicates that a more effective strategy for learning new information is now being employed by subjects in the treatment group. Parietal regions showing increased activation after treatment included superior parietal lobule, precuneus, posterior cingulate, and supramarginal gyrus. The parietal lobes have been shown to be involved with visualization and mental imagery [2831], both of which are skills that

are taught and practiced in the current treatment protocol. It was thus expected that brain regions that are active during imagery would be more active post-treatment than pretreatment due to increased engagement of these regions during the learning process. Importantly, these regions showed an increase in activation on fMRI from pre-treatment to post-treatment in the treatment group, but not in the placebo-control group (the placebo-control group was never exposed to the treatment itself). Thus, the noted changes in patterns of cerebral activation are not due to repeated testing, or changes in task, or situation novelty. The changes in patterns of cerebral activation on fMRI noted in the experimental group only can be tied to the treatment specically, as this was the only difference between the treatment group and the placebo-control group. It can be noted from Fig. 2 that the control group showed consistent (although non-signicant) decreases in activation from baseline to follow-up in every region examined. Sustained mental effort in healthy controls leads to decreased brain activity during performance, as well as over time (reviewed in [32]). This has been attributed variously to practice effects, a switch from controlled to automatic processing, priming, and habituation. The decreased activation seen in the control group in the present study (albeit a non-signicant decline) may be attributable to habituation to both the scanner and the task. As has been demonstrated in previous studies, patterns of cerebral activation are altered in MS samples relative to


J Neurol (2012) 259:13371346


healthy controls during cognitive task performance [1012, 33, 34]. Notably, this alteration is specic to individuals who show impairment in the realm assessed [10, 11]. This importantly highlights the suitability of cognitive interventions in persons with a demonstrated impairment, as cognitively intact individuals would not be expected to show much change beyond practice effects. As such, we were careful to enroll only persons with MS who had a predetermined threshold of memory decits for both the placebo-controlled group and the treatment group in the current study. The current ndings augment the literature on altered patterns of fMRI by demonstrating that improvements in cognitive function that result from a cognitive intervention with proven behavioral efcacy [9] are also evidenced on the neural level by changes in patterns of cerebral activation on fMRI. That is, the behavioral changes that have recently been documented following treatment with the mSMT can be tied directly back to increased activation in regions of the brain responsible for the tasks taught within the treatment paradigm. This study thus provides solid evidence that the mSMT results in clear changes in brain functioning that can be directly tied to the desired changes in behavior. The documentation of these neurofunctional changes following treatment with the mSMT naturally leads to additional questions that must be addressed in future studies, the rst of which questions the duration of the treatment effect. Previous behavioral studies have only examined behavioral changes post-treatment with the mSMT for 6 weeks post-treatment, noting the treatment effect to withstand the passage of 6 weeks [9]. Given that the behavioral change is sustained, one would expect that the neurofunctional change accompanying the behavioral output is also sustained, but that question remains to be addressed experimentally. Additionally, the duration of the treatment effect and techniques for maintaining the effect maximally over time are voids in the literature that can be tested via fMRI and accompanying neuropsychological evaluation. Armed with the knowledge that fMRI is a sensitive measure of changes following cognitive rehabilitation, one can also explore the application of fMRI for identifying patients that may benet more or less from the treatment. Specically, neurofunctional markers at baseline may help predict who may or may not benet from treatment. While demonstrating signicant ndings, there are methodological limitations to the present study that deserve mention. First and foremost, the current study employs a small sample. Larger samples are recommended in replication of the ndings. However, it is important to note that despite the small sample, signicant treatment effects were noted both behaviorally and on fMRI. An additional limitation is the lack of a long-term follow-up. This study was

the rst of its kind to document increases in patterns of cerebral activation post-treatment, but it did not look at the maintenance of that effect over time. This would be valuable information for both clinicians and researchers. Additionally, the current study did not systematically examine differences between those who benet from treatment and those who did not. This is an important question to ask in hopes of maximizing prescription of the treatment to those for whom benet would be expected. So, why is it important to understand patterns of cerebral activation underlying behavioral change? There are several answers to this question. First, given the fact that we were able to document regions of the brain for which increased activation is associated with improvements in verbal learning and memory, a logical next step might be to increase activation in those areas via other means (e.g., TMS, pharmacological treatment) and examine the impact on learning and memory performance. Perhaps we can supplement the mSMT to increase the activation of these targeted regions in hopes of increasing our treatment effect. Future research might also examine the effect of specic drugs identied as improving memory performance on these and other brain regions in an effort to identify a common brain region (or network of regions) for which increased activation results in better memory. Asking these and other such questions will serve to increase our knowledge of the brains response to treatment and help us focus our treatment to better achieve the desired response. The present study is the rst examining the effects of cognitive rehabilitation at a neurofunctional level in MS. We were able to demonstrate changes in brain functioning that was evident on both the neural (i.e., increased activation) and the behavioral level (i.e., improved memory performance), from before to after only ten sessions of a strictly behavioral intervention. These ndings highlight the need to further the clinical availability of the treatment as well as the provision of third party reimbursement for its use with the appropriate patients.
Acknowledgments This work was funded by National Institutes of Health grants (Grant number R01 HD045798 and HD045798-S to N.D.C.); National Multiple Sclerosis Society (training grant-MB0003 to J.D.) and the Kessler Foundation. Conicts of interest None.

1. Chelune GJ, Stott H, Pinkston J (2008) Multiple sclerosis in textbook of clinical neuropsychology. Taylor and Francis, New York 2. Chiaravalloti ND, DeLuca J (2008) Cognitive impairment in multiple sclerosis. Lancet Neurol 7(12):11391151


1346 3. Benedict RH, Zivadinov R (2006) Predicting neuropsychological abnormalities in multiple sclerosis. J Neurol Sci 245(12):6772 4. Rao SM et al (1993) Memory dysfunction in multiple sclerosis: its relation to working memory, semantic encoding and implicit learning. Neuropsychology 7(3):364374 5. DeLuca J, Barbieri-Berger S, Johnson SK (1994) The nature of memory impairments in multiple sclerosis: acquisition versus retrieval. J Clin Exp Neuropsychol 16(2):183189 6. DeLuca J et al (1998) Acquisition and storage decits in multiple sclerosis. J Clin Exp Neuropsychol 20(3):376390 7. OBrien AR et al (2008) Evidenced-based cognitive rehabilitation for persons with multiple sclerosis: a review of the literature. Arch Phys Med Rehabil 89(4):761769 8. Allen DN et al (1998) Teaching memory strategies to persons with multiple sclerosis. J Rehabil Res Dev 35(4):405410 9. Chiaravalloti ND et al (2005) Treating learning impairments improves memory performance in multiple sclerosis: a randomized clinical trial. Mult Scler 11:5868 10. Chiaravalloti ND et al (2005) Cerebral activation patterns during working memory performance in multiple sclerosis using FMRI. J Clin Exp Neuropsychol 27(1):3354 11. Mainero C et al (2004) fMRI evidence of brain reorganization during attention and memory tasks in multiple sclerosis. Neuroimage 21(3):858867 12. Staffen W et al (2002) Cognitive function and fMRI in patients with multiple sclerosis: evidence for compensatory cortical activation during an attention task. Brain 125(Pt 6):12751282 13. Filippi M et al (2010) The contribution of MRI in assessing cognitive impairment in multiple sclerosis. Neurology 75(23): 21212128 14. Jonides J (2004) How does practice make perfect? Nat Neurosci 7:1011 15. Morgen K et al (2004) Training-dependent plasticity in patients with multiple sclerosis. Brain 127(Pt 11):25062517 16. Rasova K et al (2005) Is it possible to actively and purposely make use of plasticity and adaptability in the neurorehabilitation treatment of multiple sclerosis patients? A pilot project. Clin Rehabil 19(2):170181 17. Poldrack RA et al (1998) The neural basis of visual skill learning: an fMRI study of mirror reading. Cereb Cortex 8(1):110 18. Olesen PJ, Westerberg H, Klingberg T (2004) Increased prefrontal and parietal activity after training of working memory. Nat Neurosci 7(1):7579 19. Belleville S et al (2011) Training-related brain plasticity in subjects at risk of developing Alzheimers disease. Brain 134: 16231634

J Neurol (2012) 259:13371346 20. McDonald WI et al (2001) Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 50(1):121127 21. Chiaravalloti ND, Balzano J, Moore NB, DeLuca J (2009) The open-trial selective reminding test (OT-SRT) as a tool for the assessment of learning and memory. Clin Neuropsychol 23(2): 231254 22. Cox RW (1996) AFNI: software for analysis and visualization of functional magnetic resonance neuroimages. Comput Biomed Res 29(3):162173 23. Forman SD et al (1995) Improved assessment of signicant activation in functional magnetic resonance imaging (fMRI): use of a cluster-size threshold. Magn Reson Med 33(5):636647 24. McDermott KB et al (1999) Direct comparison of episodic encoding and retrieval of words: an event-related fMRI study. Memory 7(56):661678 25. Christodoulou C et al (2006) Effects of donepezil on memory and cognition in multiple sclerosis. J Neurol Sci 245:127136 26. Krupp LB, Christodoulou C, Melvillle P et al (2004) Donepezil improved memory in multiple sclerosis in a randomized clinical trial. Neurology 63:15791585 27. Lezak MD, Howieson DB, Loring DW (eds) (2004) Neuropsychological assessment, 4th edn. Oxford University Press, New York 28. Cohen MS et al (1996) Changes in cortical activity during mental rotation. A mapping study using functional MRI. Brain 119(Pt 1): 89100 29. Formisano E et al (2002) Tracking the minds image in the brain I: time-resolved fMRI during visuospatial mental imagery. Neuron 35(1):185194 30. Knauff M et al (2000) Cortical activation evoked by visual mental imagery as measured by fMRI. Neuroreport 11(18): 39573962 31. Alivisatos BP, Petrides M (1997) Functional activation of the human brain during mental rotation. Neuropsychologia 35: 111118 32. DeLuca J et al (2008) Neural correlates of cognitive fatigue in multiple sclerosis using functional MRI. J Neurol Sci 270(12): 2839 33. Audoin B et al (2005) Magnetic resonance study of the inuence of tissue damage and cortical reorganization on PASAT performance at the earliest stage of multiple sclerosis. Hum Brain Mapp 24(3):216228 34. Audoin B et al (2003) Compensatory cortical activation observed by fMRI during a cognitive task at the earliest stage of MS. Hum Brain Mapp 20(2):5158