Computer-assisted cognitive training may help patients affected by several illnesses alleviate their cognitive deficits, or healthy people improve their mental performance. The talk presents a web-accessible system, called MS-rehab, developed specifically for the cognitive rehabilitation of patients suffering from multiple sclerosis, but usable also in other contexts. MS-rehab embeds an AI-based mechanism that adapts the difficulty of cognitive training exercises to the trainee’s performance. The AI-based mechanisms proved to be superior to a mechanism designed by medical experts in a live experiment.

1. MS-rehab: a web-accessible
system for AI-supported
cognitive training
Floriano Zini
Free University of Bozen-Bolzano
Smart Data Factory
13 Nov 2020
Mauro Gaspari
University of Bologna

2. › Multiple sclerosis (MS) is the most common autoimmune
disease affecting the central nervous system
› Cognitive impairment is observed in 43-70% of MS patients
› Impairment can affect attention, memory, and executive
functions (e.g., problem solving and reasoning)
› Negative impact on:
› learning new concepts and
methods, visual-spatial skills,
information processing abilities
› common daily and social life
activities
Cognitive impairment in MS
Verywell / JR Bee

3. › Use of computerized CR
in MS is increasing,
using ad-hoc platforms
(and videogames)
› Advantages with respect
to pen-and-paper
› Exercise variability, data collection, execution feedback,
cost optimization
› Promising results have been achieved in some
research studies
Computerized cognitive rehabilitation in MS
CogniPlus

4. › There are still doubts on the effectiveness of
computerized cognitive rehabilitation therapies
› Tools mainly dedicated to hospital
rehabilitation only, under expert guidance
› For single patients, no support for groups
› Limited support for the clinical operator in
› Configuration of rehabilitation sessions
› Monitoring
› No personalization for patients
or patient types
Limits of computerized CR in MS

5. MS-rehab: computerized CR dedicated to MS
Set up of
rehabilitation
sessions
Cognitive
profile
building
Neuropsychological
test input
CR exercise
execution
Monitoring

6. MS-rehab web architecture

7. › MS-rehab provides exercises designed by a team of
computer scientists and neuropsychologists
› The exercises cover the three main cognitive
domains:
› attention (12 exercises), memory (8), and executive
functions (4)
› each exercise has multiple versions in which the stimuli
vary, for a total of 52 exercises
› Patients can practice with simplified exercise before
starting the rehabilitation
› Patients can stop an exercise, and restart it later
› the system restarts the rehabilitation with an exercise at
the same difficulty level as the last one completed by
the patient.
› A feedback is shown at the end of each exercise
› It includes the obtained performance, and the number
of correct, wrong, and missed answers
CR exercises

8. Exercises for Attention
Selec%ve a(en%on
Tick all the target images shown on the upper
part of the screen among those visualized in
the grid below
Alterna%ng a(en%on
Select one animal and then, a;er hearing
a sound, another one, among those
sliding on the screen
Divided a(en%on
Select specific vegetable items and, at the
same =me, capture a sound whenever it is
played

9. Exercises for Memory
Memory recogni%on
Memorize the target images for some =me,
and then, a;er they disappear, recognize them
among others within a given =me
Memory face-name association
Memorize people’s names for a time interval,
and then, after they disappear, write the
correct name under each face
Memory Nback
Press the green or red button
according to whether or not the
current figure is the same as the
one appeared N figures before

10. Exercises for Executive Functions (1)
Response - inhibi%on
The pa=ent has to discriminate
between fruit and vegetables by
pressing the green or red buCon, but
should not press the green buCon if a
strawberry appears
Planning - ZooSafari visit
Plan a route through a map of the
ZooSafari, visi=ng a selec=on of
animals and loca=ons

11. Exercises for Executive Functions (2)
Planning – Day of Commitments
The patient has to find a plan that
allows her/him to execute all the tasks,
without violating the constraints, and
optimizing the number of actions
Planning – Weekend in Rome
The pa=ent has to plan a two-day
vaca=on in the Italian capital city

12. Implementation of planning exercises
› Based on AI-planning, in particular
on PDDL4J
› open-source Java toolkit that provides
state-of-the-art planners
› We exploited PDDL4J to implement
› a Scenario Generator, to generate a huge
number of solvable exercise
› a Game Engine, to verify, after any action
executed by the patient, if a plan to solve
the scenario still exists

13. MS-rehab efficacy assessment
Recruitment
of 8 MS
patients
T0: Baseline
evaluation
Treatment
with MS-
rehab
T1: Post-
treatment
evaluation
Experiment @ Lab of Cognitive Psychology Univ. of Parma
Evaluation
› Psychometric
› Cognitive/neuropsychological
(WAIS-IV scale)
Treatment
› Three weekly cycles, each including
three individual sessions of 40
minutes each
› Same session exercises for all the
patients
› Difficulty of the exercises
automatically increased Rehabilitation session

14. MS-rehab efficacy assessment
Results
Results of 2-tailed T and Wilcoxon signed-rank tests.
In bold the p-values showing a significant difference between T0 and T1 (α = .05)
WAIS-IV
VCI: Verbal Comprehension
PRI: Perceptual Reasoning
WMI: Working Memory
PSI: Processing speed
Psychometric params
BDI-II: depression
STAI-Y1, STAI-Y2: anxiety
FSS: fatigue
MSQOL-54: quality of life

15. MS-rehab vs other CR tools
Comparison of MS-rehab with state-of-the-art professional cogni7ve rehabilita7on tools
( ✔ = supported; ✘ = not supported; ● = not MS-specific)

16. From automatic exercise difficult variation …
› State-of-the-art systems for computerized
cognitive training have automatic mechanisms
to fit the difficulty of exercises to how the
subjects perform in their execution
› MS-rehab has a mechanism defined by expert
rehabilitators
› difficulty is increased when the patient performance on an
exercise overcomes the 80% of the maximum for two
consecutive times

17. … to adaptive exercise difficult variation
› The adaptability to the patient is limited by
predefined rules
› the exercise parameters are changed using a fixed policy
which is the same for all the subjects
› We have designed a truly adaptive mechanism
based on Reinforcement Learning (RL)
› The policy used to change the difficulty
level is learned while the individuals
execute the training
› Performance is used a reward indicating
the goodness of the way the difficulty
was previously modified

18. Training of 2
exercise
difficulty-
variation policies
Baseline
evaluation of
groups A and B
A and B had
sessions of the
two selected
exercises
•A used MS-rehab
embedding the policies
learned with RL
•B used the baseline
version of MS-rehab
Post-treatment
evaluation
of A and B
Adaptive mechanism assessment
Experiment
Evaluation
› PASAT test
Results
› Both A and B performed performed
significantly better in the post-
treatment evaluation
(paired t-test, p=1.7∗10-5 for group
A, and p = 0.02 for group B)
› B performed significantly better
than A (t-test)

19. Conclusions and future work
› MS-rehab is a tool dedicated to the cognitive rehabilitation of
multiple sclerosis patients, designed in tight cooperation with
clinical experts
› fully integrated within the MS cognitive rehabilitation process
› with coverage of the cognitive domains that are affected by MS
› with advanced monitoring of rehabilitation activities
› embedding AI modules for that drive planning exercises and adaptive
exercise difficult variation
› Further experimentation is needed to prove its clinical
efficacy
› A system reengineering would be needed to improve
system robustness

20. Acknowledgements
› Dr. Sergio Stecchi for contributing to the design of MS-rehab with
his medical expertise
› Dr. Debora Castellano, Dr. Federica Pinardi, Dr. Francesca Rizzi, Dr.
Fabio Bellomi, Dr. Beatrice Goretti, Dr. Federica Lato, Dr. Enrico
Montanari and Dr. Livia Ludovico for their help in the analysis of the
MS cognitive rehabilitation process
› Dr. Elena Maria Bressan, Dr. Daniele Baschieri, Dr. Margherita
Donnici, and Dr. Bartolomeo Lombardi for the contribution given to
the developement of MS-rehab with their master thesis.
› Laboratory of Cognitive Psychology of the Department of Medicine
and Surgery of the University of Parma, and Prof. Olimpia Pino and
Dr. Ciro Urselli for their contribution to the pilot study of MS-rehab
› Franca Stablum and Agnieszka Kolasińska of the Department of
General Psychology, University of Padua for extensively testing MS-
rehab
The MS-rehab project was partially supported by Fondazione Del Monte
and Delia Gennari ONLUS.

21. Thank you for your attention!
MS-rehab website:
https://rehab.cs.unibo.it/MS-rehab-website
Contact:
Floriano Zini
Free University of Bozen-Bolzano - Smart Data Factory
floriano.zini@unibz.it
smart@unibz.it
https://smart.inf.unibz.it/