People with severe cerebral palsy (CP) who are nonverbal and unable to control conventional computer interfaces due to the severe limitations in hand control benefit from eye-tracking technology as access method to Augmentative and Alternative Communication (AAC) devices and to computers for education and leisure. Research has put forward the large demands that the use of AAC puts on working memory (WM), defined as our ability to temporarily store information that is no longer perceptually present, allowing us to manipulate it for meaningful goal-directed behaviour.

People with CP show significant WM deficits, which affect learning capacities and academic achievement, including impaired language and reading comprehension, and arithmetic difficulties. Cogmed WM training (CWMT) is a computerized software with a great potential to boost WM capacity and overall cognitive functioning. Its effectiveness is influenced by the theory of neuroplasticity due to repeated mental tasks.

To date, no prior study investigated the effectiveness of CWMT in children and youth with severe CP who rely on eye-tracking technology for daily-life functioning. This is the first trial that aims to explore the impact of a 5-week CWMT on WM capacity and its near-transfer effect (trained and untrained WM tasks), far-transfer effect (other cognitive abilities, quality of eye movements and behaviour) and retention 3-months post intervention.

Working memory (WM) is one of the core executive functions (EF) which encompass higher-order cognitive abilities, critical for optimal daily-life functioning. WM is defined as our ability to temporarily store information that is no longer perceptually present, allowing us to manipulate it for meaningful goal-directed behaviors, such as decision-making, problem-solving and reasoning. WM includes verbal WM and (nonverbal) visuospatial WM, two main components defined by content. Verbal WM is responsible for temporarily storing verbal information such as letters, words, numbers or nameable objects, and is a strong predictor of language development and reading comprehension. Visuo-spatial WM is a fundamental component of the eye movement system, visual perceptual functioning and it is a strong predictor of arithmetic performance. WM is related to active long-term memory aiding the retrieval of stored information from past experiences to successfully execute the task at present. WM is practically involved in most (if not all) daily-life activities, from simple tasks like remembering a phone number or following instructions, to highly complex tasks associated with learning, academic achievements in literacy, numeracy and science, or language comprehension.

One of the central limitations of human cognition is the restricted WM capacity, i.e., the amount and duration of information that can be stored and manipulated at once. Low WM capacity affects approximately 15% of all children, from which over 80% are at very high risk of educational underachievement. In terms of its neural basis, prior work consistently put forward an important role of fronto-parietal networks in WM performance with some variation depending on stimulus type. In typically developing (TD) people, WM continues developing throughout childhood and peaks in adolescence along with a number of structural maturational processes in the brain. Significant WM deficits have been described in an array of neurodevelopmental disorders.

Cerebral palsy (CP) is the most common cause of severe physical disability in childhood with a prevalence of 1.6-3.4 per 1000 livebirths. It comprises a group of developmental impairments of movement and posture attributed to nonprogressive lesions in the developing fetal or infant brain. CP is categorized into spastic, dyskinetic, ataxic and mixed forms, and functioning of individuals ranges from mild to severe levels of limitations. In CP, 41% of all people are not able to walk independently, 23% are unable to handle objects using their hands, and 32% are nonverbal. The motor impairments are frequently accompanied by impairments of cognition and behaviour. EF in CP are underreported to date, particularly in the non-ambulant and non-verbal cases of CP where these impairments tend to be overlooked due to challenges in finding appropriate assessment tools that do not rely on motor and verbal skills. Namely, only 36.8% of children with severe motor impairments in CP have their cognitive functioning assessed, compared to 96.5% of children with mild CP. Both verbal and visuospatial WM deficits seem to be present in all forms of CP and more so as severity of functional limitations increases. In CP, WM deficits are associated to impaired reading comprehension and arithmetic difficulties. Learning difficulties are present in around 40% and visual-perceptual impairments in 40-50% of all people with CP.

In addition to central neural alterations, children with CP also show altered functioning of the autonomic nervous system (ANS), as evidenced by inherent reduced heart rate variability (HRV). In TD population, higher performance of WM (and EFs in general) is strongly associated with increased HRV, an established marker of parasympathetic ANS drive (facilitating 'rest and digest'). This suggests that HRV also forms an important marker indexing effective cognitive function. In TD, increased HRV is accompanied by increased resting-state functional brain connectivity. Investigations into the relative contribution of HRV differences in CP and their relation to EF deficiencies and to core functional connectivity networks previously identified as related to WM have not been addressed in any prior study.

People with CP with severe motor impairments benefit from eye-tracking technology as an access method to assistive and alternative communication devices (AAC) and to computers for education and leisure. Eye movements are tracked by an infrared sensor and translated to cursor movements on the screen by which children can navigate and select icons of interest. Nonverbal children with CP have significantly impaired language comprehension, especially prominent in spastic CP. Eye-tracking technology leads to increased communication outcomes, and it has a positive impact on quality of life through increased activity levels, participation levels, self-efficacy, and self-esteem. With its undeniable benefits in mind, the use of AAC puts large demands on WM. For example, to express wants and needs, these children need to remember the symbol of interest, all the while navigating through an array of symbols not all of which are simultaneously present, remembering the most efficient path, inhibiting possible distractions and finally locating and selecting the target symbol. The use of AAC systems can be a real challenge for children with low WM capacity, likely leading to technology discontinuance, and loss of opportunities and benefits. People with CP also show lower quality of eye movements compared to TD people, however, eye movement accuracy can be improved through intensive eye-tracking training. In addition, eye-tracking methods have been previously used to successfully index WM, a method which has not yet been investigated in people with CP, but may hold important clinical implications.

Computerized WM training emerged as a novel, non-invasive treatment option with great potential to boost WM capacity and overall cognitive functioning. EF share capacity constraints and have overlapping neural systems, which explains the generalized effect of WM training to other cognitive functions. A recent systematic review synthesized the existing evidence about cognitive interventions in CP, highlighting, among others, the low quality of existing evidence in WM training programs and the need for future, more robust studies in CP. Previous research in cognitive interventions (in general) spans from activity interventions to boost cognitive abilities (i.e. climbing, dancing, hippotherapy) to specific training programs, such as Mi-Yoga mindfulness-based movement program, Move-It-to-improve-it (Mitii), and Cogmed WM Training (CWMT). Due to the severe motor impairments, the goal group of severe CP (GMFCS levels IV-V) are not able to perform the activity interventions of climbing, dancing, hippotherapy. Furthermore, they are not able to perform the movement-based programs, such as the Mi-Yoga and Mitii, leaving the CWMT program as the most feasible training program for children and youth with severe CP who can access their computers using eye-tracking devices. Other computerized intervention training programs include SMART (in the developing stages, not specific to WM training thereby not useful to meet the project's goals), and NeuronUp (https://www.neuronup.com/; not specific to WM training but rather of executive functions in general), for which there is currently only a published study protocol of an RCT in mild CP, however, the authors have not yet published any results so there is no evidence to support its use to boost cognitive abilities in general, and of WM in specific.

CWMT is currently the most used training software with the most empirical evidence available on its effectiveness in different populations. The software itself relies on adaptive training using algorithms, that is, the difficulty of the WM tasks is constantly kept at the highest to challenge WM capacity but still fit to the abilities of each user. WM Training is strongly influenced by the theory of neuroplasticity which relies on repeated and tailored training, hence the adaptive algorithm. Neuroplasticity changes have been reported after in different populations as well as in premature children (including few participants with CP). In spite of extensive research using CWMT, the general conclusion is that the increased WM capacity as a result of the training has repeatedly shown near-transfer effects but inconclusive far-transfer and retention effects, the latter largely dependent on age, population, training type and duration. CWMT is the best WM training software to use in this project for various reasons: (1) it is designed specifically to boost WM capacity, which is not the case for the other computerized training programs that aim to improve executive functions in general (including cognitive flexibility, attention etc.); this is a critical point to consider as the main goal of this project is to assess any changes primarily in WM capacity, and then if any skills are transferred to the other executive functions; (2) CWMT program has substantially more evidence for its positive impact on WM capacity explored in high quality studies, (3) training is structured with 25 sessions in total, and tasks within each session are adapted to the participant's abilities, offering a targeted and customized training experience which is crucial given the heterogeneity that characterizes CP as a diagnosis; (4) CWMT sessions can be performed using eye-tracking devices, which is critical for the goal group of children and youth with severe CP.

The effectiveness of WM Training in CP remains largely unexplored. To the researchers knowledge, the only study that used CWMT (as a training program with the most empirical evidence to date to improve WM capacity) in children with spastic CP reported significant improvements post-training in both near-transfer and far-transfer skills, including in visuo-spatial skills, inhibition and phonological processing. Two studies using the CWMT in extremely-low and very-low birth-weight children (each study included two participants with CP) also reported both near- and far-transfer effects, retained at 6-months follow-up. Although with promising results, these studies excluded children with CP with severe motor impairments, leaving a gap in knowledge of importance to be explored.

Inclusion Criteria:

1. official CP diagnosis by a paediatric neurologist
2. 7-21 years old
3. users of eye-tracking technology for computer access and AAC
4. classified as level IV-V on the Manual Ability Classification System (MACS)
5. classified as level I-III on the Eye-pointing Classification Scale (EpCS)
6. ability to understand and follow instructions, assessed using the Dichotomous Choice Screen

Exclusion Criteria:

1. severe visual and/or hearing impairment
2. presence of photosensitive epilepsy

• De-identification may not be sufficient to ensure participant confidentiality.
• The informed consent form did not include an option for data sharing on clinicaltrials.gov