Cognitive Rehabilitation in Multiple Sclerosis

Cognitive Rehabilitation in Multiple Sclerosis: Comparison

Please note this is a comparison between Version 1 by BRUNO BROCHET and Version 4 by Lindsay Dong.

Cognitive rehabilitation (CR) is the most promising approach for treating multiple sclerosis (MS)-related cognitive impairment (CI), despite important methodological shortcomings. CR programs could include techniques designed to improve specific domains of cognitive function such as memory, attention, or executive functions, but they can also include psychotherapy targeting emotional symptoms, behavioral interventions, and interventions targeting psychomotor issues such as motor–cognitive interference.

multiple sclerosis
cognitive rehabilitation
cognition
memory

1. Introduction

Treating cognitive impairment (CI) in multiple sclerosis (MS), the leading cause of disability due to nontraumatic neurological disease in young adults, is an important challenge [1]. The contribution of CI to disability in MS has been increasingly recognized, and CI has been shown to decrease health-related quality of life (HR-QoL), even in the early stages of the disease [2]. CI negatively affects daily activities such as driving, vocational status, absenteeism, and instrumental activities in persons living with MS [3]. No medication has proven to have a consistent symptomatic effect on CI in MS, and disease-modifying therapies only have a limited impact on CI progression [4].

CI in MS is dominated by a slowdown in information processing speed (IPS), as well as by disturbances of more specific cognitive functions such as attention, episodic memory (EM), working memory (WM), and executive function (EF) [5]. If a relatively circumscribed alteration in IPS linked to a specific process deficit can occur, changes in IPS can alter other cognitive processes and usually reflect cognitive functioning and efficiency. The alteration of IPS has consequences for WM, attention, EF, and EM. IPS impairment predicts later disability vocational status and changes in quality of life [3].

Cognitive rehabilitation (CR) is the most promising approach for treating MS-related CI, despite important methodological shortcomings. CR programs could include techniques designed to improve specific domains of cognitive function such as memory, attention, or executive functions, but they can also include psychotherapy targeting emotional symptoms, behavioral interventions, and interventions targeting psychomotor issues such as motor–cognitive interference.

A wide variety of neuropsychological tests have been used in the assessment of cognitive impairment in MS. A panel of experts recently proposed recommendations [3] for screening tests—the Symbol Digit Modalities Test (SDMT) [6], the Processing Speed Test (PST) [7], and the Computerized Speed Cognitive Test (CSCT) [8], a self-report scale (Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ) [9]—and recommended brief cognitive batteries: the Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS) [10], the Brief Repeatable Neuropsychological Battery (BRNB) [11], and the Minimal Assessment of Cognitive Function in MS (MACFIMS) [12]. While most of the CR studies used the SDMT, the other assessment tests varied widely from study to study.

2. Cognitive Rehabilitation in Multiple Sclerosis

2.1. Individual-Specific Rehabilitation

3.1. Individual-Specific Rehabilitation

2.1.1. Memory

3.1.1. Memory

After an initial pilot study dedicated to the modified-story memory technique (mSMT)-based rehabilitation method ^[13][41], which is basically an imagery- and context-based memory retraining program, Chiaravalloti et al. [14] conducted a pivotal study on a larger sample (86 patients) with a positive result on the main criterion, the learning slope of an NP test of EM (California Verbal Learning Test-second edition, CVLT-II), and a positive effect on Functional Assessment of MS (FAMS), an assessment of HR-QoL, as a secondary endpoint. These positive results were maintained at the remote evaluation performed six months after treatment. The second primary endpoint assessing everyday objective memory (RBMT) was also significantly more improved in the treated group than in the control group but with a small effect size. This study did not show any efficacy of booster sessions.

Assuming that cognitive training focused on improving new learning via memory-enhancing techniques does not transfer to tasks other than those involved in the intervention and considering that items self-generated by an individual are better remembered than provided information, another method of memory rehabilitation was studied by this group, the self-generation learning program (Self-GEN) ^[15]. In this program, the therapist teaches patients to discover the benefit of using the self-generation strategy during training, and participants must discuss how the strategies could be used in other tasks. The study involved 35 MS patients and reached primary endpoints (the Contextual Memory Test, an objective learning and memory test, and the Self-Regulation Skills Interview that assesses self-awareness and strategy use). Prospective memory (Memory of Intention Test) and HR-QoL (FAMS) were also improved. However, no significant differences were observed between groups for the CVLT-II, the MFQ, and Actual Reality, which is an ecological assessment of functional performances during an internet-based task. Therefore, the hypothesis that this rehabilitation technique could be transferred to other cognitive functions, notably the executive functions, could not be confirmed.

Assuming that cognitive training focused on improving new learning via memory-enhancing techniques does not transfer to tasks other than those involved in the intervention and considering that items self-generated by an individual are better remembered than provided information, another method of memory rehabilitation was studied by this group, the self-generation learning program (Self-GEN) [27]. In this program, the therapist teaches patients to discover the benefit of using the self-generation strategy during training, and participants must discuss how the strategies could be used in other tasks. The study involved 35 MS patients and reached primary endpoints (the Contextual Memory Test, an objective learning and memory test, and the Self-Regulation Skills Interview that assesses self-awareness and strategy use). Prospective memory (Memory of Intention Test) and HR-QoL (FAMS) were also improved. However, no significant differences were observed between groups for the CVLT-II, the MFQ, and Actual Reality, which is an ecological assessment of functional performances during an internet-based task. Therefore, the hypothesis that this rehabilitation technique could be transferred to other cognitive functions, notably the executive functions, could not be confirmed.

In a recent pilot study, Self-GEN strategy was associated with two other strategies of memory and learning rehabilitation, space learning (SL), and retrieval practice (RP), in the Strategy-based Training to Enhance Memory (STEM) program ^[16]. This pilot study, with limited power (20 patients), did not show a significant superiority of the STEM program to the control intervention (non-training program), on the primary endpoint (CVLT-II) and other objective tests of memory (BVMT-R), but it did show some encouraging trends and positive results on HR-QoL (FAMS and Short-Form 36 [SF-36]).

In a recent pilot study, Self-GEN strategy was associated with two other strategies of memory and learning rehabilitation, space learning (SL), and retrieval practice (RP), in the Strategy-based Training to Enhance Memory (STEM) program [38]. This pilot study, with limited power (20 patients), did not show a significant superiority of the STEM program to the control intervention (non-training program), on the primary endpoint (CVLT-II) and other objective tests of memory (BVMT-R), but it did show some encouraging trends and positive results on HR-QoL (FAMS and Short-Form 36 [SF-36]).

2.1.2. Multiple Domains

3.1.2. Multiple Domains

A study compared a 13-week individual NP rehabilitation program in a large sample of patients (99) with no intervention ^[17][18][16,40]. The primary endpoint of this study was the subjective perception of deficits using the Perceived Deficits Questionnaire (PDQ). A greater improvement of this score was observed in the rehabilitated group compared to the control group. Regarding the NP tests, such as the Symbol Digit Modalities Test (SDMT), which was the second primary outcome, and the other tests of the Brief-Repeatable Battery (BRB), no significant improvement was observed, except for the Trial Making Test-A (TMT-A). Note that the intervention’s total duration in this partially negative study was much lower than in the other two individual multidomain studies (780 min versus 3600 and 2250 min).

The REACTIV study was launched to demonstrate the superiority of a specific CR program (REACTIV) over nonspecific intervention (NSI) for NP assessment, virtual reality cognitive testing, and daily cognitive functioning in MS ^[19][34]. It was a single-blind RCT comparing these two interventions in patients with MS selected based on CI at specific tests of IPS, WM, and EF. Both programs included 50 individual sessions administered three times a week for 17 weeks. The specific intervention was tailored to patients’ deficits. The primary endpoint was NP assessment of IPS, attention, EF, and WM. Secondary endpoints included ecological assessment by tasks in a virtual reality environment (Urban Daily Cog^®) and daily cognitive functioning assessment. More NP scores improved significantly in the active group and several NP scores, alertness and divided attention, and the ecological assessments improved significantly more after specific CR than after NSI. Lastly, SCR improved daily cognitive functioning. Most improvements were maintained 4 months after the end of the intervention. However, HR-Qol was not shown to be improved. The study showed the interest of an individualized and intense intervention including a meta-cognitive approach. It was also the first to show a transfer in ecological tasks. However, the study was underpowered for showing a larger effect on cognitive function. The study was performed in a real-world setting with rehabilitation by speech therapists in city practice.

2.2. Group Interventions

3.2. Group Interventions

A double-blind multicenter study compared group rehabilitation to a sham intervention (non-training) ^[20][32]. In a fairly large sample of patients selected on the basis of the presence of a cognitive deficit, the main evaluation criterion was EM measured by the Selective Reminding Test (SRT), and a significant difference was observed in the learning curve between the treated group and the control group. WM also improved. However, there was no improvement in the other functions studied (EF, IPS). There was also no positive effect on HR-QoL.

The last study is difficult to interpret because of a mixed intervention combining group management and CT using Lumosity software ^[21]. Moreover, patients could be included regardless of the presence or absence of CI, and the percentage of patients who were actually cognitively impaired at inclusion is not reported. Overall, the study was negative, the primary endpoint, the Everyday Problems Test—Revised, a self-administered questionnaire, was negative, and among the NP tests, only two were improved: the CVLT and the PASAT. This negative outcome could possibly be due to a significant improvement in the control group treated by a computer game.

2.3. Computerized Multidomain Rehabilitation

The last study is difficult to interpret because of a mixed intervention combining group management and CT using Lumosity software [30]. Moreover, patients could be included regardless of the presence or absence of CI, and the percentage of patients who were actually cognitively impaired at inclusion is not reported. Overall, the study was negative, the primary endpoint, the Everyday Problems Test—Revised, a self-administered questionnaire, was negative, and among the NP tests, only two were improved: the CVLT and the PASAT. This negative outcome could possibly be due to a significant improvement in the control group treated by a computer game.

3.3. Computerized Multidomain Rehabilitation

The RCTs concerning multidomain CR, including attention, IPS, EF, and WM, have been developed in three categories of setting: in institutions, at home with telerehabilitation (online), and at home offline. The duration and number of CT rehabilitation sessions varied from one study to another with sessions lasting 30 to 60 min with a frequency of 2 to 3 per week for a duration of 4 to 12 weeks and a total duration of 300 to 2160 min.

Different software programs were used, the most frequent being RehaCom^®. Up to 2021, eight studies have been published with this software, and four blinded RCTs were selected for this review ^{[22][23][24][25]}[15,22,25,36]. Several studies have looked at other CT programs in MS and are summarized in Table 15.

Table 15. Blinded RCT of CT in MS.

Session Duration (Min)

CR Duration (Weeks)

Total Duration (Min)	Control Intervention	Main Positive Results (Primary Outcomes)	LTFU (Mths)	Positive Results (Secondary Outcomes)	Other Positive Results	Main Negative Results (Primary Outcomes Underlined)
Cerasa et al. [15]	2013	RehaCom	60 min twice a week	6	720	Placebo training	no specified primary outcome	ND		ST	BRB tests, TMT
Amato et al. [17]	2014	APT	60 min twice a week	12	1440	Placebo training	no specified primary outcome	6		PASAT	SDMT *
De Giglio et al. [18]	2015	DKBT	30 min 3 times a week	8	720	No intervention	no specified primary outcome	ND		ST, SDMT, some MSQoL54 subscales	PASAT
Hancock et al. [19]	2015	PS Insight and Brain Twister n-back	30 min 6 times a week	6	1080	Sham training	PASAT	ND			SDMT, LNS, Digit backward
Pedullà et al. [21]	2016	Cogni-Track	30 min 5 times a week	8	1200	Non-adaptative training	no specified primary outcome	6		SRT, SDMT, PASAT, WLG	SPART, WCST
Campbell et al. [22]	2016	RehaCom	45 min 3 times a week	6	810	Placebo training	SDMT	3			BVMT, CVLT, FAMS, EQ-5D
Charvet et al. [23]	2017	Brain HQ (PS)	60 min 5 times a week	12	1800	Placebo training	Composite NP score	ND
Grasso et al. [24]	2017	APT	60 min 3 times a week	12	2160	No intervention	no specified primary outcome	6	**
Messinis et al. [25]	2017	RehaCom	60 min twice a week	10	1200	No intervention	no specified primary outcome	6		Verbal EM, attention, verbal fluency, IPS z scores
Chiaravalloti et al. [26]	2018	SPT	30 min twice a week	5	300	No intervention		ND	TIADL		WAIS digit symbol *
Messinis et al. [36]	2020	RehaCom	45 min 3 times a week	8	1080	Placebo training	SDMT, GVLT, BVMT	ND	EQ-5D
Vilou et al. [37]	2020	Brain HQ (PS)	40 min twice a week	6	480	No intervention	no specified primary outcome	ND		GVLT, BVMT, TMT-A, ST	SDMT
Blair et al. [39]	2021	Cogmed	30–45 min 5 times a week	5	750–1125	No intervention		6			PASAT , SDMT , ST *

1st Author

Year

Intervention (Software)

Session Duration (Min)

CR Duration (Weeks)	Total Duration (Min)	Control Intervention	Main Positive Results (Primary Outcomes)	LTFU (Mths)	Positive Results (Secondary Outcomes)	Other Positive Results	Main Negative Results (Primary Outcomes Underlined)
Cerasa et al. ^[22]	2013	RehaCom	60 min twice a week	6	720	Placebo training	no specified primary outcome	ND		ST	BRB tests, TMT
Amato et al. ^[26]	2014	APT	60 min twice a week	12	1440	Placebo training	no specified primary outcome	6		PASAT	SDMT *
De Giglio et al. ^[27]	2015	DKBT	30 min 3 times a week	8	720	No intervention	no specified primary outcome	ND		ST, SDMT, some MSQoL54 subscales	PASAT
Hancock et al. ^[28]	2015	PS Insight and Brain Twister n-back	30 min 6 times a week	6	1080	Sham training	PASAT	ND			SDMT, LNS, Digit backward
Pedullà et al. ^[29]	2016	Cogni-Track	30 min 5 times a week	8	1200	Non-adaptative training	no specified primary outcome	6		SRT, SDMT, PASAT, WLG	SPART, WCST
Campbell et al. ^[23]	2016	RehaCom	45 min 3 times a week	6	810	Placebo training	SDMT	3			BVMT, CVLT, FAMS, EQ-5D
Charvet et al. ^[30]	2017	Brain HQ (PS)	60 min 5 times a week	12	1800	Placebo training	Composite NP score	ND
Grasso et al. ^[31]	2017	APT	60 min 3 times a week	12	2160	No intervention	no specified primary outcome	6	**
Messinis et al. ^[24]	2017	RehaCom	60 min twice a week	10	1200	No intervention	no specified primary outcome	6		Verbal EM, attention, verbal fluency, IPS z scores
Chiaravalloti et al. ^[32]	2018	SPT	30 min twice a week	5	300	No intervention		ND	TIADL		WAIS digit symbol *
Messinis et al. ^[25]	2020	RehaCom	45 min 3 times a week	8	1080	Placebo training	SDMT, GVLT, BVMT	ND	EQ-5D
Vilou et al. ^[33]	2020	Brain HQ (PS)	40 min twice a week	6	480	No intervention	no specified primary outcome	ND		GVLT, BVMT, TMT-A, ST	SDMT
Blair et al. ^[34]	2021	Cogmed	30–45 min 5 times a week	5	750–1125	No intervention		6			PASAT , SDMT , ST *

CR: cognitive rehabilitation; Min: minutes; LTFU: positive results at long-term follow-up; ND: not done; Mths: months; EM: episodic memory; IPS: information processing speed; *: test improved in both groups; ** no difference between groups. Rehabilitation programs: ATP: Attention Processing Training program; DKBT: Dr. Kawashima’s Brain Training

Blinded RCT of CT in MS.

1st Author	Year	Intervention (Software)

^®

(Nintendo); PS: Posit Science; SPT: Speed of Processing Training. Outcomes: BRB: Brief-Repeatable Battery; BVMT-R: Brief Visuospatial Memory Test—Revised; CVLT-II: California Verbal Learning Test—2nd edition; EQ-5D: EuroQoL five-dimension questionnaire; FAMS: Functional Assessment of Multiple Sclerosis; GVLT: Greek Verbal Learning Test; LNS: Letter–Number Sequencing; MSQoL54: MS Quality of Life questionnaire; PASAT: Paced Auditory Serial Addition Test; SDMT: Symbol Digit Modalities Test; SPART: Spatial Recall Test (10–36); SRT: Selective Reminding Test; ST: Stroop test; TIADL: Timed Instrumental Activities of Daily Living Test; TMT: Trail-Making Test. WAIS: Wechsler Adult Intelligence Scale; WCST: Wisconsin Card-Sorting Test; WLG: Word List Generation. Underlined outcomes are primary outcomes.

3. Conclusions

Altogether, the studies about memory individual CR support the efficacy of m-SMT for improving learning in MS patients. This technique is able to improve verbal episodic memory and HR-QoL. There is some evidence of the efficiency of daily cognitive functioning, but there is no evidence of transfer in other domains. More data are necessary to support other methods of memory individual CR.

Most studies of specialized individual rehabilitation have given positive and encouraging results, but this type of rehabilitation requires significant human resources and a significant time investment by patients.

The large number of studies about CT is probably due to the ease of this technique. The possibility of implementing this CT at home and with telerehabilitation is also very promising. The variety of software used in these studies makes it difficult to compare them. RehaCom has been the most used. All in all, these studies on RehaCom tend to show the effectiveness of this rehabilitation program in MS but leave several important questions unanswered: its use in telerehabilitation, its effectiveness on specific sub-domains according to the cognitive domain most affected, the interest of adding individual rehabilitation sessions including in particular a meta-cognitive approach, its effectiveness on daily cognitive functioning assessed by ecological tests, and its effect on quality of life in RRMS.

The other studies of CT with different programs (APT, PS, SPT) showed various results. As a result of the different software used, it is difficult to draw general conclusions, but some of them demonstrated the feasibility of home-based CR ^[26][27][29] and of telerehabilitation ^[30][33][34]. The evidence of the efficacy of telerehabilitation is still limited, but the results encourage further studies in this field. The development of computerized assessment methods for cognitive disorders in MS, which have been the subject of a recent review ^[35], may help the deployment of telerehabilitation. Although a demonstration of the effectiveness of multidomain CR, in particular on quality of life and transfer to cognitive domains different from those trained, is lacking, several studies show the feasibility of individual CR in daily life settings and home telerehabilitation using CT. A certain level of efficacy has been established regarding NP outcomes and in several studies on daily cognitive functioning.

Although a demonstration of the effectiveness of multidomain CR, in particular on quality of life and transfer to cognitive domains different from those trained, is lacking, several studies show the feasibility of individual CR in daily life settings and home telerehabilitation using CT. A certain level of efficacy has been established regarding NP outcomes and in several studies on daily cognitive functioning.