Constraint-Induced Movement therapy for long-term walking impairment in multiple sclerosis

Our research laboratory at the University of Alabama at Birmingham (UAB) in the United States has tested a distinguished form of physical therapy for persons with chronic walking difficulty from multiple sclerosis. The therapy is called Constraint-Induced Movement therapy, or CI therapy for short.  The treatment was developed from years of basic neuroscience research by Professor Edward Taub at UAB and then transferred to persons with stroke starting about 20 years ago.

CI therapy is designed to improve the amount and quality of use of an impaired arm or both legs in the real world, outside of the clinic, where the effects of rehabilitation matter the most. The treatment has the following components:

  1.  intensive practice with the impaired part of the body on real-world-like tasks under a therapist’s supervision,
  2. shaping the training exercises, so that persons move from simpler to more difficult levels of accomplishment,
  3. constraint from being passive or relying on compensatory activities, and
  4. a program of behavioral techniques to transfer to gains from the clinic to the real-life situation.

This program of behavioral techniques is called the “transfer package” and includes:

  • signing a behavioral contract between the client and therapist, in which the client promises to do the assigned activities and review any problems with the therapist;
  • practicing assigned exercises at home each day after therapy;
  • maintaining a home practice diary;
  • reporting on the quality of leg use each day;
  • and a daily, half-hour problem-solving session with the therapist, to review how practice at home went and to resolve any difficulties with home practice.

Although many of these techniques are used in standard physical therapy, no other program uses all of these together. Previous research by our laboratory not only demonstrated that the transfer package improves gains from CI therapy in stroke patients compared to treatment without the transfer package, but also that it expands brain areas involved with movement and sensation as measured from standard MRI scans, unlike treatment without the transfer package.

Our current research enrolled 4 individuals with chronic multiple sclerosis who had difficulty with using their legs because of the disease. We gave each of them 3.5 hours/ day of CI therapy over 3 consecutive weeks (excluding weekends), thus for 52.5 hours each. Training exercises included partial body-weight supported treadmill training, stair climbing, and stepping over obstacles. Rest periods were given as often as needed. Our main outcome measure was the Lower Extremity Motor Activity Log, or LE-MAL. The LE-MAL is an interview concerning the quality of real-world leg use on 14 common activities, such as walking outdoors, turning while standing, and getting in and out of bed.

The results were that all of our patients significantly improved on the LE-MAL, not only at the end of treatment, but also at least 1 year afterward without further treatment. After the first year, one patient passed away because of an unrelated illness, one patient became worse in his leg control to the point where he returned to his baseline, while the other two patients maintained their gains for as much as 4 years of follow-up. No patients suffered any ill effects from the treatment. Fatigue levels slightly improved as well.

Our findings are unprecedented for rehabilitation for multiple sclerosis: no other treatment study thus far has followed persons with MS for more than 6 months following treatment, nor has demonstrated improvement on real-world activities for at least 1 year afterward. Although one of our patients lost the improvement after 1 year, we believe that our findings thus far suggest that CI therapy can promote improved leg use in the community for at least 6 months for most individuals, and possibly much longer. Furthermore, for persons who lose their gains after 1 year, it may be possible to improve their function again with a repeat course of treatment.

Most important is that this therapy is most appropriate for persons with multiple sclerosis who are in the chronic, non-relapsing phase of illness. In this phase, there is little help from approved medications, and rehabilitation efforts have otherwise not demonstrated the transfer of gains from the clinic to the real world. CI therapy cannot restore a person to completely normal function. However, we believe that it can improve the amount and quality of spontaneous use of the legs in the community for a very long time without the need for repeated therapy or additional medications.

We plan further studies that will enroll larger numbers of participants, so that we can understand better how consistent are these initially promising findings. Future research should also examine whether persons who lose their gains after CI therapy can improve again with a repeat course of treatment.

About Victor Mark

Victor MarkVictor Mark, MD, is a neurologist in the Department of Physical Medicine and Rehabilitation for the University of Alabama at Birmingham (UAB) in the United States. He was specially trained in the field of Behavioral Neurology, and brings to his work expertise with behavioral influences on recovery and rehabilitation after different neurological diseases, including stroke, multiple sclerosis, spinal cord injury, and traumatic brain injury. He has been the Medical Director for the CI Therapy research and clinical programs at UAB for the past 14 years.  Any questions about his work can be directed to him at


Mark VW, Taub E, Uswatte G, Bashir K, Cutter GR, Bryson CC, Bishop-McKay S, & Bowman MH (2013). Constraint-induced movement therapy for the lower extremities in multiple sclerosis: case series with 4-year follow-up. Archives of Phys Med Rehab, 94 (4), 753-60 PMID: 23111280

To receive the PDF of the full article please contact the author Victor Mark



  1. Regarding screening for lower extremity CI therapy, at present the inclusion criteria are kept minimal. In our clinic, we will generally treat any person with a chronic neurologic disorder who indicates having impaired ambulation and without marked pain. These individuals are nearly always self-referred. On the LE-MAL (Lower Extremity Motor Activity Log), in our stroke clinic, from the first 89 persons who were treated (over a span of about 10 years; each person was treated for 3 consecutive weeks), the mean LE-MAL score was 4.6, range 0.5-8.4. The mean improvement at post-treatment was by 1.8 points (standard deviation = 1.0). Only 1 person worsened or stayed the same (entry LE-MAL score = 5.2, post-treatment score 5.0). The mean percentage improvement across the 89 persons was 39% on the LE-MAL.
    We also require that they indicate that they are capable of walking at least 16 meters 5x/ day with or without an assistive device, but without the aid of another person. This is to assure that persons have the stamina to participate, and that their ambulation is self-initiated.

    For our research, we require that their LE-MAL score to be no greater than 6.5 (on a scale whose top score = 10). This is so that we can avoid encountering ceiling effects when assessing treatment change that would otherwise invalidate statistical analysis. As further persons become treated, we will learn whether these criteria should be adjusted.
    Otherwise, no, we do not have specific screening criteria for LE CI therapy. In patients who come from far away, we ask either that they send to us a short video recording to show how well they can walk in the home, or more recently we have changed to using Skype to allow our therapists to evaluate individuals remotely, so that persons whom we feel would not be capable of treatment do not have to travel to see us.

    Once accepted patients arrive at our clinic, I conduct a screening medical evaluation to assure general medical stability, and our therapists conduct sample training tasks to be sure that persons can comply with tasks. They also evaluate passive range of movement in the legs to check for possible limitations of movement. After that, we do not conduct further screening. This approach has generally worked well. On some occasions the therapists will ask that I write a prescription for an ankle-foot orthosis, if the patient either does not have one or needs a new one, depending on their judgment.

  2. stuart miller says:

    Dr. Mark, thanks again for your feedback. Just one more question on screening for appropriate patients for CIMT for LE. Other than low LE -MAL scores to indicate decreased or non-use, is there a screening test (similar to ‘tennis-ball’ test for UE) that would indicate an appropriate patient for treatment intervention ?

  3. stuart miller says:

    Dr. Mark, thanks again for your feedback. Just one more question on screening for appropriate patients for CIMT for LE. Other than low LE -MAL scores to indicate decreased or non-use, is there a screening test (like ‘tennis-ball’ test) that would indicate an appropriate patient for treatment intervention ?

  4. Thank you for your important questions.
    1) Our article indicates that fatigue scores did improve in our patients, but not to a significant extent for the group as a whole. With a treatment sample of only 4 persons, it would not be possible to say that a change would be significant. The calculated Effect Size however was 0.9, which was large, but it would be best to await larger trials of CI therapy before drawing conclusions.
    It should be noted, however, that many studies of other forms of physical therapy for MS have reported improved fatigue ratings. I therefore believe that improved fatigue is a common result of physical therapy, which is very gratifying.
    Our participants did not have neuropathic pain. We anticipate that as we study larger patient samples, we may encounter this. We generally allow patients with pain scale ratings of 4/10 or less to participate, but will always terminate treatment for any individual who finds treatment uncomfortable. In our roughly 20 years of treating persons with various neurological disorders, we have seldom encountered pain, and patients who had mild pain asked to remain in treatment because they appreciated their improved movement.
    CI therapy basically works with whatever tools are available. The tasks to be used are decided by agreement between the client and the therapist. What we feel matters most is (1) selecting tasks that the client can perform, (2) tasks that the client would agree to perform, and (3) including a diversity of tasks to maintain interest. In our lower extremity work, we include activities such as ascending/ descending stairs, balance exercises while standing (including bending from the waist to retrieve objects at lower levels), stepping over small obstacles, and walking with progressively narrower paths, besides treadmill-assisted training. Although we have not tested the approach, I believe that working without a treadmill should be as successful as including it, as long as the principles of shaping, positive reinforcement, feedback, and the “transfer package” of behavioral reinforcement principles are followed (described briefly in our present article; described in greater detail in Morris DM, Taub E, Mark VW. Constraint-Induced Movement therapy (CI therapy): characterizing the intervention protocol. Eura Medicophys 2006;42:257-268).

  5. stuart miller says:

    Dr. Mark, thanks for your insight. I recognize that there was a significant improvement on the LE-MAL scores which is great. Was there also an improvement in overall self-reported fatigue levels with the patients and finally was there an improvement in neuropathic pain (or was this an issue with any of the patients) ? I look forward to your larger study. Thanks again. For those without the body-weight supported treadmill system, do you have recommendations ?

  6. Victor Mark, MD says:

    A final point also is that if we can demonstrate consistently beneficial results of lower extremity CI theray in a larger group of individuals with MS, using a randomized clinical triall-design, then we would feel ready to train other therapists in our methods, as we do currently with the arm for of CI therapy. The latter is taught in our biannual training course for therapists that is provided in Birmingham, USA, details of which can be found at

  7. Victor Mark, MD says:

    Thank you for your interest in our research and your well-considered questions. With regard to the length of treatment for arm vs leg training in CI therapy, our work with stroke patients, extending back into the late 1990s, observed that individuals who underwent leg training tended to plateau in the 3rd week, whereas individuals who underwent arm training leveled off in the 2nd week. Furthermore, greater gains were seen in our leg-trained patients in the 3rd week before leveling off, compared to the earlier weeks. We unfortunately do not have similar studies in persons with MS, having trained far fewer of them. We hope that if our work can be funded to extend our initial observations to a larger group of individuals, we could reassess how many weeks of treatment are best to obtain maximal improvement during direct training.

    For the treatment environment, we did not make any specific adjustments. Our participants were trained in the basement of our academic hospital, which may have helped with providing the ideal temperature for training (although, to be sure, I have not checked the thermostat there lately, but have always found this to be comfortable). Our participants never complained of heat-related discomfort during treatment.

    Finally, the question about a modified CIMT approach is very valuable. Some investigators have shown that stretching out the same amount of contact hours over several more weeks to months is associated with results the same as giving the treatment over just 2 weeks, for persons with arm training and stroke. In our first article on giving CI therapy to persons with MS (published in Multiple Sclerosis in 2008), we actually had our initial participants get treatment for their arms between 2 weeks and 10 weeks. Overall, the results were very positive, but too few individuals were entered into the various different treatment schedules on which to base a conclusion.

    Extending the treatment over multiple weeks-months could be one way to accommodate the different availabilities of participants for this form of training. We gave CI therapy for as much as 10 weeks in our first research study because of difficulty with attending our treatment on consecutive days. For the present, we prefer to treat everyone on the same, 2 or 3 week schedule, to allow consistency for scientific control. This is however hard for some persons to attend.

    It is indeed possible that a more extended form of CI therapy could be very beneficial and thus accommodate the different availabilities of persons with MS (many of whom, as you know, remain employed). However, one problem with this approach is that the longer that treatment is extended, the greater the chance either of persons dropping out due to vacations, moving away, or other changes, or scheduling difficulties because of staff-related changes (vacations, etc).

    At this time we do not have the necessary information pertaining to the optimal dosing of CI therapy–how many hours, how many days? We hope that if we can continue to investigate our current regimen and find consistent improvement, then we will have the basis to explore the results of adjusting the dosage.

  8. Dr. Mark, first of all, brilliant work to apply to MS population what has such strong evidence for stroke patients with non-use. I am curious why you did 3 weeks of intensive work vs 2 weeks as is common with CVA. As well, other than rest breaks provided whenever needed, was there any other alteration in the environment (? cooler) while doing CIMT ? Do you feel that a modified CIMT approach may also be helpful ? Just looking for ways to transfer it easily to the clinic setting. Thanks a lot for this extremely helpful study.