PEMOCS: theory derivation of a concept for PErsonalized MOtor-Cognitive exergame training in chronic Stroke—a methodological paper with an application example

Background Coping with residual cognitive and gait impairments is a prominent unmet need in community-dwelling chronic stroke survivors. Motor-cognitive exergames may be promising to address this unmet need. However, many studies have so far implemented motor-cognitive exergame interventions in an unstructured manner and suitable application protocols remain yet unclear. We, therefore, aimed to summarize existing literature on this topic, and developed a training concept for motor-cognitive exergame interventions in chronic stroke. Methods The development of the training concept for personalized motor-cognitive exergame training for stroke (PEMOCS) followed Theory Derivation procedures. This comprised (1.1) a thorough (narrative) literature search on long-term stroke rehabilitation; (1.2) a wider literature search beyond the topic of interest to identify analogies, and to induce creativity; (2) the identification of parent theories; (3) the adoption of suitable content or structure of the main parent theory; and (4) the induction of modifications to adapt it to the new field of interest. We also considered several aspects of the “Framework for Developing and Evaluating Complex Interventions” by the Medical Research Council. Specifically, a feasibility study was conducted, and refining actions based on the findings were performed. Results A training concept for improving cognitive functions and gait in community-dwelling chronic stroke survivors should consider the principles for neuroplasticity, (motor) skill learning, and training. We suggest using a step-based exergame training for at least 12 weeks, 2–3 times a week for approximately 45 min. Gentile's Taxonomy for Motor Learning was identified as suitable fundament for the personalized progression and variability rules, and extended by a third cognitive dimension. Concepts and models from related fields inspired further additions and modifications to the concept. Conclusion We propose the PEMOCS concept for improving cognitive functioning and gait in community-dwelling chronic stroke survivors, which serves as a guide for structuring and implementing motor-cognitive exergame interventions. Future research should focus on developing objective performance parameters that enable personalized progression independent of the chosen exergame type.


S.2) Training Dosage
Based on the recommendations described in 'D.2) Defining the Training Dosage', the PEMOCS concept will be applied with the following training principles (Table S1); in the planned RCT, participants will train twice a week, 30-40 minutes over 12 weeks, resulting in a total training volume of 840 minutes.The following practical aspects played a role.Twelve weeks was assumed a feasible duration for the study intervention, especially when striving for an acceptable recruitment rate.Regarding frequency, we found in our feasibility study that two sessions per week were feasible for an on-site training intervention at rehabilitation centres, sessions of 30-40 minutes were most accepted (Huber et al. 2021).Time progression will be applied by adding two minutes of training every week (Table S1).

S.3) Progression Rules
The "Dividat Senso" provides a selection of games, which target different cognitive functions.Most games train several functions, however, for all of them a main target function can be defined, which can be categorized into attention (AF), working memory (WM), executive functions (EF), memory (MF), and visuospatial functions (VSF) (for a description of all games, see Supplement 2).AF and WM were fused into one cognitive sub-dimension out of two reasons, (1) the task measuring working memory in the MoCA is counted in the domain of AF (Nasreddine et al. 2005), which plays a role as the MoCA is used to allocate the domains to the extended taxonomy (see below, section S.3.1).( 2) Allocating the working memory game to the AF sub-dimensions helped filling the motor-cognitive skill categories in a more balanced way (see Figure S1, Nomis = WM game).Therefore, these four cognitive sub-domains (AF/WM, EF, MF, VSF) were chosen for the implementation of the PEMOCS concept (Table S2).

S.3.1) Personalizing the Allocation of the Cognitive Sub-dimensions
To allocate the four cognitive domains to "least" to "most" impaired [see 'D.S2).This delivers a score between 0 and 100 for each test, which represents how well the participant performed in the test compared to the norm sample, and also categorizes them into 'below average' (< 16), 'average' (16-84), and 'above average' (> 84).By comparing the scores of the different tests, the order of the domains from "least" to "most" impaired is delivered.S2: (Nasreddine et al. 2005, Freitas et al. 2012, Coen et al. 2016)

S.3.2) Assigning Motor-cognitive Tasks to the Sub-dimensions of the Extended Taxonomy
The different motor-cognitive tasks within this exergaming system are on the one hand determined by the different "Dividat Senso" games, which target different cognitive functions (Supplement 2).On the other hand, various motor tasks can be executed on the pressure sensitive plate, e.g.steps, squats, running movements and jumps (Table S3).To categorize these motor tasks, they were first assigned to one of the two sub-dimensions of 'Action Function' (Figure S1) by the following rules; 'Body stability' includes tasks, where the centre of pressure (COP) moves within stable or slightly and slowly moving limits of stability.'Body transport', on the other hand, includes tasks, where the COP moves within constantly and rapidly moving limits of stability.Second, the motor tasks within the two sub-dimensions (Body stability, Body transport) are ranked in ascending difficulty and allocated to three motor-task levels a, b and c (Table S3).
By combining the different games and motor tasks, different versions of all games were derived.
Possible combinations were determined based on the game mechanics, e.g.some games require directed stepping and are therefore not applicable for combination with body-weight shifting.Each game version was then assigned to a motor-cognitive skill category in the extended taxonomy (Figure S1).To do so, the allocation rules presented in 'D.3.2) Assigning Motor-cognitive Tasks to the Sub-Dimensions of the Extended Taxonomy' were applied as follows: • Stationary: games with a still game scene In-motion: games with a moving game scene • No Intertrial Variability: stimuli appear in a fixed sequence and speed Intertrial Variability: stimuli appear in a random sequence and / or speed • Body Stability: games, which are played with a "stability" motor task (see Table S4) Body Transport: games, which are played with a "transport" motor task (see

S.3.3) Progression from Session to Session
The difficulty levels 1 and 2 as well as 6 and 7 were fused to provide enough variability for the training sessions as in these levels, several motor-cognitive skill categories in the extended taxonomy remained empty (Figure S1).Therefore, the following five difficulty levels exist in this application example: • 1-2: red / orange, including motor-skill categories 1A, 1B, 2A • 3: yellow, including motor-skill categories 1C, 2B, 3A • 4: green, including motor-skill categories 1D, 2C, 3B, 4A • 5: blue, including motor-skill categories 2D, 3C, 4B • 6-7: purple / grey, including motor-skill categories 3D, 4C, 4D All participants start in difficulty level 1-2 (first training session is standardized) and stay there for both trainings in week 1.From the second session on, progression steps to subsequent sessions are determined using the procedure presented in 'D.3.3)Progression from session to session'.Participants rate their perceived performance and perceived motor-cognitive task difficulty using the visual analogue scales (Figure 2) in each session.As objective evaluation of challenge, the training supervisor notes their perception of how optimally challenged the participant was on a scale of -2 (strongly overchallenged) to +2 (strongly underchallenged, Figure S2).This S-score (Figure S3) represents the OP-score in this application example [see Figure 3  provide progression over the blocks, the motor tasks levels are used (Table S3, Figure S4).The motor tasks are applied as follows (compare Figure S4): • Warm-up: any motor task that serves the purpose of re-familiarization • First block: motor-task level a • Middle block: motor-task level c • Last block: motor-task level b (b or c if there is no middle block) • Cool-down: any motor task that fosters a feeling of success The progression within the blocks ensures the focus on the most impaired cognitive domain.Based on the session time (Table S1), the time for each block and game was defined (Table S4).A game should last between 90 and 180 seconds, therefore, in weeks 1 and 2, there are two blocks, while in weeks 3 to 12, there are three blocks.In each block, the game targeting the most impaired domain (IV) lasts 180s.The remaining time of the block is distributed equally to the games targeting the less impaired domains (I-III, Table S4).The games are arranged in each block as defined in 'D.3.4)Progression within Sessions' (Figures 3 & S4).
• In case a game should be played > 2x / session (e.g. because it is the only one of a domain in the current sub-level), it is allowed to replace it once by another game from a lower difficulty level.• In level 6-7, the warm-up and cool-down games can also be chosen from difficulty level 4 (instead only from difficulty level 5).• For participants, who stay in level 6-7 for at least 4 sessions and cannot progress further, two games in two different domains per session can be exchanged with one of the following games from sublevel 4.4: Nomis_2C, Flexi_2C, Shop_2C, Simon_2C.Game settings at least as difficult as in sublevel 4.4 should be used, e.g.longer sequences or shopping lists or Flexi_2C can be complemented with an additional calculation task.• Besides these rules, game preferences by the participants can be considered.

Refining Steps after the Feasibility Study
Besides the steps, which are described in the main document, two further problems had been detected in the feasibility study, which directly relate to the application as presented here.(1) Several cognitive functions, namely working memory, memory, and visuospatial functions were under-represented in the list of motor-cognitive activities that were provided by the exergaming system "Dividat Senso" (Huber et al. 2021).Therefore, several new games were developed for the upcoming RCT.This helped ensuring the ability to focus on all of the tailored cognitive domains, and filling more motor-cognitive skill categories [see 'E.3) Fill empty motor-cognitive skill categories'].(2) We found in the feasibility study (Huber et al. 2021), that the motor-cognitive task difficulty had been below the targeted range optimal for motor learning and training in chronic stroke (Akizuki and Ohashi 2015).We, therefore, integrated new motor tasks, new games and game settings with higher nominal task difficulty into the concept (Table S2, for a more detailed description of the motor tasks, see Supplement 2).With the larger selection of motor-cognitive tasks, especially with higher nominal task difficulty, the PEMOCS concept should now provide sufficient functional task difficulty and intensity for a wide range of chronic stroke survivors.

Limitations
The presented application example of the PEMOCS concept renders several limitations.First, the applied training variables range at the lower border of the presented recommendations [compare 'D.2) Training Dosage'].However, as this application example will be used in a clinical trial, it also had to fulfil feasibility criteria.Based on the experience in the feasibility study (Huber et al. 2021), we suspected that a longer study duration, a higher frequency, or longer sessions would decrease the recruitment rate of the RCT.Nevertheless, the planned training variables still move within the recommended range.(2) Despite our effort to fill as many motor-cognitive skill categories as possible, several still remained empty (Figure S1).This was due to the selection of available games, and some limitations in the possibility to adjust game mechanics.We accounted for this by fusing the levels 1 and 2, as well as 6 and 7 [section 'S.3.3)Progression from Session to Session'], by introducing the sub-levels, and by adding additional variability rules for the levels with many empty categories [section 'S.4) Variability Rules'].(3) The objective evaluation of challenge in this application example is the supervisor's perception of how strongly the participant was challenged.This is not as objective as it should optimally be (compare 'Limitations and Future directions' in the main article).Optimal would be, if the exergaming system would deliver one or several performance parameters, which would allow the determination of whether the participant has reached a performance threshold that implies a progression or retrogression to the next level.This was not possible with the used system, and no other performance parameter was yet available (see 'Future Directions and Limitations' in the main article).We mentioned the need for research on feasible performance parameters already in our feasibility study (Huber et al. 2021).Unfortunately, this was not possible in the time since, as more research on different performance parameters and their indication for optimal functional task difficulty is needed.Therefore, supervisor's ratings will be used.Subjectivity of the rating is reduced by instructions for the supervisors on how to evaluate the performance of the participant as objectively as possible, and by them noting their evaluation without knowing that the participant's rating will be.
(Nasreddine et al. 2005)tion of the cognitive domains'], the MoCA [Montreal Cognitive Assessment,(Nasreddine et al. 2005)] is used.Achieved percentage values of the four domains in the MoCA at baseline are ordered; the lowest is assigned to IV, the second lowest to III, and so forth.To confirm the assignment or in case two domains rank equally, further cognitive baseline assessments are considered.

Table S4 )
• No Object Manipulation: participant cannot manipulate (move) the objects in the game Object Manipulation: participant can manipulate (move) an object in the game • Cognitive Dimension: Main target cognitive functions of the games were grouped to the four sub-dimensions, attentional functions (AF) and working memory (WM), executive functions (EF), memory functions (MF), and visuospatial functions (VSF), see Supplement 2.

Table S3 :
Motor (Wuest et al. 2014)cation example of the PEMOCS concept.The tasks are assigned to "Body Stability" or "Body Transport" and categorized into motor-task levels a (easiest) to c (most difficult).For a more detailed description of the motor tasks, see Supplement 2, TableS6.BWS; body-weight shifts.B-pad; soft balance pad used as an unstable surface, which is placed on the plate.Supplementary FigureS1: Extended taxonomy with an example how the cognitive domains could be allocation and the different games versions in the motor-cognitive skill categories.AF / WM: attention and working memory, EF: executive functions, MF: memory, VSF: visuospatial functions.Colours: Difficulty levels based on(Wuest et al. 2014); red / orange: level 1-2; yellow: level 3; green: level 4; blue: level 5; purple / grey: level 6-7.Game names: numbers (e.g.Simple_1A) represent environmental sub-dimension of the game version, letters (e.g.Simple_1A) the action function sub-dimension.