Keynote:

Beyond the Research: The Things I've Learned in My Academic Career

Biography:

Krissy Doyle-Thomas is an administrator, professor, and researcher at Mohawk College, where she played a pivotal role in co-developing and now leads Ontario’s first graduate certificate programs in brain disorders management and mental health & disability management.

She earned her PhD in Medical Sciences from McMaster University and completed a Post-Doctoral Fellowship with Dr. Evdokia Anagnostou at the Autism Research Centre at Holland Bloorview Kids Rehabilitation Hospital. During this fellowship, she used Magnetic Resonance Imaging (MRI) to investigate the structural and functional properties of the brain in individuals with Autism Spectrum Disorders and their connection to clinical symptoms.

Doyle-Thomas also held key research administrative roles at Holland Bloorview Kids Rehabilitation Hospital. These roles included manager of research operations for the Bloorview Research Institute, as well as research associate positions with the Autism Research Centre (under Dr. Anagnostou) and the Communication Neuroscience Discovery Team (under Dr. Deryk Beal).

Her commitment to applying research to improve health practices for marginalized communities is evident in her current work. She is focused on enhancing mental health support for racialized communities and increasing research participation among underrepresented groups to better tailor health services. Funded by the Canadian Institutes of Health Research (CIHR), she collaborates with community-based health professionals and academic institutions to ensure evidence-based practices that meet the specific needs of these communities.

Doyle-Thomas’s contributions have been recognized with several accolades: in 2024, she was honoured as a Woman of Influence by the RBC Canadian Women Entrepreneur Awards; in 2022, she was named one of the 100 Accomplished Black Canadian Women in the 4th edition of the book by the same name; and in 2017, she was featured in CBC News' "HERStory In Black" documentary.

University and Program: Queen’s University- Honours Bachelor of Health Sciences

Supervisor and Lab: Dr. Melanie Penner, Accept Lab - Autism Research Centre

Project Title: A Comprehensive Characterization of the Extensive Needs Service Pilot Program Participants in Ontario

Principal Investigator: Dr. Melanie Penner

Research Summary:

The study aims to describe the baseline demographics and health measures of participants in Ontario's Extensive Needs Service (ENS) pilot program which is designed to address unmet needs due to developmental, mental, and social complexities. Eighteen families from Holland Bloorview were assessed using their demographics, Brief Family Distress Scale (BFDS), Pediatric Global Health-7 (PGH-7), and Behaviour Assessment System for Children-3 (BASC-3) scores. This categorization aims to optimize ENS program delivery and evaluation to better serve the community.

Research Question: What are the baseline demographics and health measures of the ENS participants?

Methods & Analyses: Participants' demographics, BFDS, PGH-7, and BASC-3 scores were analyzed using proportions, medians, means, and standard deviations.

Results: Half of the participants were aged 6-11, with 37.5% being racial minorities. Median BFDS and PGH-7 scores indicated significant family distress and fair mental health. Mean BASC-3 T-scores for externalizing problems were 67.8, suggesting at-risk levels.

Conclusions: Baseline data reveals significant family distress and poor mental health among ENS participants. These insights will help refine the ENS program by improving intake processes and tailored program delivery to better address the needs of youth with complex issues.

Relevance to Holland Bloorview: By categorizing and describing ENS families, the program can better meet distinct individual needs. This will also help in targeting advocacy for these populations whose needs are being left unmet by the system, ultimately benefiting Holland Bloorview clients and families.

Bio: 

Bisola is a third-year Queen’s University student pursuing a BHSc in Health Sciences. She is deeply interested in the value of holistic and accessible healthcare, arising from her lived experience back home in Nigeria. She observed various root causes of health problems, their impact on individual and community health as well as a lack of plausible interventions. After moving to Canada, she realized that these issues also existed here, and decided to become a health advocate. Working as a research trainee in the Autism Research Centre has opened her eyes to knowledge gaps in developmental disability research and introduced her to the ongoing work to bridge these gaps and support affected populations. Working with Dr. Melanie Penner at the ACCEPT lab has been invaluable and the learning opportunities have led to her growth and expansion of interests. This experience has also fueled her aspirations of becoming a family physician and a clinician scientist.

University and Program: McMaster University, Health Sciences

Supervisor and Principal Investigator: Dr. De-Lawrence Lamptey

Project Title: Investigating the predictors of academic success among minority-identifying and Indigenous children during COVID-19

Research Summary: 

Childhood disability remains inadequately addressed in research, particularly in the context of unprecedented disruptions such as the COVID-19 pandemic. While the generalized effects of the pandemic are beginning to be understood, the specific impacts on the academic success of racialized children with disabilities in Canada have received less attention. Rectifying the gaps in contemporary research is a key step towards adequately supporting the distinct educational needs of minority children with disabilities post-pandemic.

Research Question: 

Is there a relationship between childhood disability and academic success?

Methods & Analyses: 

A crowdsourcing questionnaire was completed by 32,000 parents of children aged 0-14 to collect information about family concerns and activities during COVID-19. A binary regression was used to identify correlations between academic outcomes and select socioeconomic factors, such as minority status, immigration status, and disability status.

Results: 

Preliminary analyses demonstrated statistically significant correlations between minority/Indigenous/immigrant status, disability status, and resultant higher concern levels for academic success and higher frequencies of structured academic activities.

Conclusions: 

While there is a demonstrated correlation between minority/Indigenous status, disability status, and negative academic outcomes, further probability sample-based research is essential to draw more meaningful and universal conclusions.

Relevance to Holland Bloorview: 

Understanding key predictors of academic success during COVID-19 can prompt further research and guide personalized educational support for children with disabilities, ensuring targeted interventions and strategies that enhance learning outcomes. This proactive approach can empower educators and caregivers to tailor educational experiences that maximize the potential of every child, particularly those from minority/Indigenous backgrounds or with disabilities.

Bio:

Chi is a second-year student at McMaster University, pursuing a Bachelor of Health Sciences degree. He is deeply interested in exploring the application of innovative technologies to enhance social welfare, specifically in the context of healthcare accessibility and patient outcomes. In his first summer as a Ward Summer Student, Chi is working in the EMBARK Lab - led by Dr. De Lawrence Lamptey - to examine whether ‘minority-identifying children with disabilities’ fared worse academic outcomes than their non-disabled peers. He driven to address the health disparities that exist within marginalized communities - his own and others alike. His ultimate goal is to use his knowledge to drive change, whether big or small, within the healthcare system, to get people the support they need and deserve.

University and Program: University of Toronto, Temerty Faculty of Medicine

Supervisors and Lab: Dr. Evdokia Anagnostou and Dr. Jacob Ellegood, Autism Research Centre

Project Title: Early Identification of Autism Spectrum Disorder using XGBoost and Convolutional Neural Networks (CNN) with Magnetic Resonance Imaging (MRI) Data

Principal Investigator: Dr. Evdokia Anagnostou

Research Summary: 

Autism Spectrum Disorder (ASD) is a highly heterogeneous, neurodevelopmental disorder characterized by challenges in social communication and restrictive, repetitive behaviours or interests. One in 50 children in Canada are diagnosed with ASD. Unfortunately, wait-times for ASD diagnosis and services are 1-3 years in Ontario. Resting-state functional MRI (rs-fMRI) shows great promise in diagnosing ASD, identifying altered brain connectivity by age 2 compared to controls. Similarly, structural MRI (sMRI) identifies regional brain differences in ASD at ages 2-5. While children ages 5-18 can be predicted with ASD using MRI, XGBoost and CNN, few studies have developed models for children ages 2-5.

Research Question: 

How effective are age-specific XGBoost and CNN models in detecting ASD in children ages 2-18?

Methods & Analyses: 

CNN and XGBoost models classified 738 ASD from 597 TD and evaluated for accuracy, AUC and F1-Score.

Results: 

XGBoost model using sMRI achieved an average 73% prediction accuracy, 0.71 AUC and 0.74 F1-score for classifying ASD from TD in pre-school children. Temporal and frontal lobe regions contributed many features with high Shapley values towards model predictions. CNN prediction accuracy for ASD using rs-fMRI is expected to exceed 70% based on age-related differences in functional connectivity.

Conclusions: 

Age-specific machine learning models using MRI may help overcome the complex heterogeneity in ASD and serve as future tools for reducing wait-times for diagnosis. However, image acquisition and data processing barriers are real-world challenges to identifying reliable biological markers for ASD and developing clinical decision-supports systems that support their care.

Relevance to Holland Bloorview: 

Early diagnosis of ASD is critical in facilitating timely access to individualized supports and therapies to improve quality of life. This project may help refine knowledge on the development of computer-assisted instruments for enhancing early screening and personalized interventions for young children with neurodevelopmental conditions. 

Bio:

Elliott Wong is a third-year medical student at the University of Toronto (UofT). Growing up, he learned to care for his special-needs sister who requires around-the-clock care. This gave rise to his passion to develop solutions for individuals with complex health needs. To that end, he completed a MEng at UofT where he learned to develop analytical and image segmentation algorithms. Elliott is currently learning to develop neural networks to guide diagnosis/interventions for children with neurodevelopmental conditions. He hopes to become a physician who cares for special-needs individuals, underserved patients and contributes towards technologies that enhance their quality of life.

University and Program: McMaster University, Mechanical and Biomedical Engineering

Supervisor and Lab: Jan Andrysek, PROPEL Lab

Project Title: Development and Validation of Machine Learning Algorithms to Evaluate Overall Walking Patterns of Lower Limb Prosthetic Users using Inertial Sensors

Principal Investigator: Jan Andrysek

Research Summary:

Objective gait evaluation can help inform rehabilitation goals for lower limb prosthetic users (LLPUs). Summarized gait quality measures enable simple interpretation and monitoring of gait patterns but traditionally require complex motion capture systems that are not widely accessible. Wearable technology serves as a promising alternative in gait assessment to encourage participation beyond clinical settings and provide better insights about real-world walking. However, robust algorithms are required for translating wearable sensor data into clinically meaningful gait metrics.

Research Question: 

Validate a machine learning (ML) algorithm to assess changes in gait patterns corresponding to clinically relevant gait parameters for LLPUs.

Methods & Analyses: 

14 LLPUs were instrumented with inertial sensors to collect gyroscope and accelerometer data during walking trials. Gait patterns were altered using an auditory biofeedback system developed by the lab. 3 ML algorithms (hidden Markov model, self-organizing map, dynamic time warping) were implemented to assess validity and sensitivity based on changes to clinically relevant gait parameters (stance-time symmetry, step length, etc.). Assessments were performed using regression and effect size statistical methods.

Results: 

Preliminary investigations have shown that changes in gait temporal symmetry correspond to changes in ML-based similarity scores using data from just 1-2 inertial sensors.

Conclusions: 

Easily interpretable assessment using wearable sensors can improve access to gait monitoring. Next steps involve integrating and testing in real-world settings.

Relevance to Holland Bloorview: 

Wearable sensors can provide simple gait tracking within and outside of the clinic. This can support clinical decision-making and encourage clients to take an active role in their rehabilitation and care.

Bio: 

Emilie is a fourth-year Mechanical and Biomedical Engineering student at McMaster University, working in the PROPEL Lab under the supervision of Dr. Jan Andrysek. She is passionate about the intersection of technology and healthcare for improving accessibility and quality of care. This summer, Emilie is working alongside PhD candidate Gabriel Ng to assess machine learning algorithms used for generalized gait quality monitoring. This project has fueled her interest in the characterization of human gait and its clinical applications in the field of rehabilitation. Emilie has deeply appreciated the learning opportunities and mentorship she has received as a Ward Summer Student.

University and Program: York University - Honours Biomedical Science

Supervisor and Lab: Dr. Tom Chau, Jenny Tou - PRISM Lab

Project Title: Creating and Testing a Brain-Controlled Music Instrument for Therapy in Children with Cerebral Palsy

Principal Investigator: Dr. Tom Chau

Research Summary:

Children with cerebral palsy (CP) often experience variations in attention span, which can hinder other areas of cognitive development. The limitations of current interventions prompted this study to explore Brain-Computer Interface (BCI) and Neurologic Music Therapy (NMT) as potential solutions. BCI offers children with communication challenges a means to control devices and interfaces using their brain activity, while NMT employs musical activities to stimulate neural pathways associated with attention and motor skills. Extensive research has demonstrated NMT’s effectiveness in enhancing attention and concentration in children with neurological disorders. Therefore, integrating BCI into NMT interventions could improve accessibility for children with CP, fostering therapeutic engagement and cognitive development.

Research Question: 

Can an EEG-BCI-enabled instrument be used in neurologic music therapy to improve attention in children with cerebral palsy?

Methods & Analyses: 

The BCI R-Net headset and music sound beam were configured using a relay box and the Mindset application to give children with CP control of the music instrument through mental commands.

Results: 

The project yielded conclusive results demonstrating the successful configuration of the BCI R-Net headset with the music sound beam via a relay box. System testing substantiated that BCI integration facilitated user control of the sound beam, producing a musical output through mental commands.

Conclusion: 

Integrating BCI technology into music therapy offers an innovative and enjoyable approach to enhance therapeutic outcomes for children with CP. The BCI-enabled music instrument developed through this project can be utilized in further studies and rehabilitation interventions to enhance cognitive development in children with neurodevelopmental disorders.

Relevance to Holland Bloorview: 

Incorporating EEG-BCI technology in music therapy offers engaging and effective interventions, enhancing the quality of life for children with CP and their families. This research underscores Holland Bloorview's commitment to pioneering advancements through holistic, family-centered care, ensuring that treatment is not only effective, but is supportive and enjoyable for all those involved.

Bio: 

Huda has recently graduated with an Honours Bachelor of Biomedical Science Degree from York University. She is passionate about the intersectional aspect of neurorehabilitation and assistive technology. In particular, Huda is interested in exploring the clinical applications of brain-computer interface (BCI) technology as a means to improve the quality of life for individuals with disabilities. During the WARD summer program, Huda designed a BCI-controlled music instrument in collaboration with other researchers, clinicians, and music therapists at Holland Bloorview. Following the completion of the summer program, Huda will conduct a case study to analyze the effectiveness of using the BCI-controlled music instrument in music therapy for children with disabilities.

University and Program: McMaster University – Health Sciences – Child Health Specialization

Supervisor and Lab: Dr. Sally Lindsay, TRAIL Lab

Project Title: Understanding recommendations to policy, practice, and removal of barriers to enhance the experiences of youth with disabilities from minoritized groups

Principal Investigator: Dr. Sally Lindsay

Research Summary:

Youth and young adults living with disabilities experience varying forms of discrimination, often exacerbated by prejudices against their respective marginalized groups, such as racialized groups. While recommendations are often considered from the perspective of decision-makers, clinicians, or related stakeholders involved in client care, it is crucial to consider the perspectives of youth living with disabilities.

Research Question: 

What shifts to policy and practice may be needed, and barriers removed to ensure that youth with disabilities thrive in all their diversity? 

Methods & Analyses: 

A qualitative interview study that involved a purposive sample of youth and young adults with disabilities, identifying as belonging to a racial, ethnic and/or gender minority group.

Results: 

Youth with disabilities belonging to minoritized groups seek practices and policies that foster more inclusive spaces that are readily accessible and continued advocacy for disability awareness.

Conclusions: 

Continued research must focus on implementing a client-centered approach to initiatives that aim to address relevant recommendations.

Relevance to Holland Bloorview: 

By highlighting lived experiences from the perspective of clients, shifts and barriers that are especially impactful can be more specifically targeted or removed to reduce harm, allowing for a safer and less encumbered transition into adulthood, alleviating burdens experienced by families and clients.

Bio: 

Jaden has been a volunteer at Holland Bloorview since 2019 and is now working in the TRAIL Lab under the guidance of Dr. Sally Lindsay. From the Ronald McDonald Playroom to The Independence Program, Jaden has immersed himself into the plethora of volunteering programs offered at Holland Bloorview (and absolutely loves it!). He continues demonstrating his passion for working with children through his research this past summer into the experiences of youth with disabilities from minoritized groups and is grateful for all the support from the TRAIL Lab team and the Ward Summer Program.

University and Program: McMaster University – Integrated Biomedical Engineering and Health Sciences

Supervisor and Lab: Dr. Tom Chau and Erica Floreani, PRISM Lab

Project Title: Deep Learning Enabled EEG Artifact Removal Algorithms for Real-World BCI Applications

Principal Investigator: Dr. Tom Chau

Research Summary:

In brain computer interfaces (BCIs), scalp EEG is used to extract instructions and allow the user to perform motor tasks through an external device. However, artifacts such as eye blinks and jaw movements often contaminate EEG signals and obfuscate meaningful physiological information. Existing artifact-removal techniques cannot readily be applied in real-time, limiting their practicality for BCIs. Deep Learning (DL) algorithms can be efficiently deployed and can advantageously learn complex relationships in data, making them a promising new method for EEG artifact removal.

Research Question: 

Does EEG artifact removal using DL techniques improve performance, as measured by classification accuracy, over traditional methods in real-world BCI tasks?

Methods & Analyses:

A pipeline that can preprocess, train, and analyze data was constructed. Existing DL models from literature and one novel model were trained and evaluated (training loss, validation accuracy, correlation between cleaned/contaminated samples) on artifact-contaminated EEG data from an open-access dataset, EEGdenoiseNet. The trained DL models were then used to remove artifacts from real-world BCI task data from a second open-source dataset. Classification accuracy of the BCI task before and after DL-artifact removal were compared and benchmarked against traditional artifact removal methods. 

Results: 

In the novel model, classification accuracy of the BCI task improved after DL-artifact removal in both the low (0.436±0.027 vs. 0.543±0.042) and high (0.702±0.040 vs. 0.776±0.048) performer groups.

Conclusions: 

This study offers preliminary evidence of the DL model’s potential in accurately removing artifacts from real-world BCI datasets. Future work will focus on quantifying the noise level of the datasets and improving the DL model architecture.

Relevance to Holland Bloorview: 

Development of novel EEG artifact removal techniques will enhance BCI usability in real-world environments, extending benefits to children with severe motor impairments.

Bio: 

Jerry is a fourth-year student at McMaster University, pursuing a B.Eng. in Engineering and Biomedical Engineering with a specialization in Mechatronics. He is deeply interested in exploring the applications of deep learning algorithms to enhance Brain Computer Interface (BCI) tasks. As a Ward Summer Student, Jerry is working alongside PhD candidate Erica Floreani in the PRISM lab, led by Dr. Tom Chau, to design novel EEG artifact removal algorithms with deep learning and develop a pipeline for evaluating the different algorithms’ performance in real-world BCI tasks. His project enabled him to develop a better understanding of the state-of-the-art techniques behind EEG artifact removal, clinical applications of BCI, deep learning in healthcare, and biomedical signal processing. Jerry has previously contributed to research involving Florescent Guided Imaging with a focus on quantifying fluorescence concentration in post resection residual cancer with deep learning. Jerry has a deep appreciation for the many learning opportunities involved during his work in the PRISM lab.

University and Program: University of Toronto Mississauga - Honours Bachelor of Science - Biology for Health Science and Sociology

Supervisor and Principal Investigator: Dr. De-Lawrence Lamptey

Project Title: Parental Concerns on Patience and Behaviour Management for Children with Disabilities During COVID-19

Research Summary: 

During the COVID-19 pandemic, the changes to the healthcare system were drastic and sudden. The goal of this research is to investigate the impact of a child's disability on parental concerns regarding having patience and managing their child's behaviour and emotions during the COVID-19 pandemic. By analyzing crowdsourcing data from Canadian parents, the study aims to identify significant demographic factors that influence these concerns. The findings will provide insights into the most vulnerable families, highlighting the need for targeted support and resources to help these families navigate the unique stresses brought about by the pandemic.

Research Question: 

How does a child’s disability affect parental concerns about patience and behaviour/emotional management during the COVID-19 pandemic?

Methods & Analyses: 

“Impacts of COVID-19 on Canadians - Parenting During the Pandemic” crowdsourcing data collected by Statistics Canada was analyzed. Binomial regression models were used to assess the family demographic factors associated with these concerns.

Results: 

Parents of children with disabilities reported significantly higher concerns about managing behaviour during COVID-19. Child disability status (X² = 2.728, p < .001) and employment status of all adults working from home (X² = 1.201, p = .001) were significant predictors.

Conclusions: 

Disability status significantly affects parental concerns about managing children's behaviour during the pandemic, highlighting the need for targeted support and equitable access to these resources.

Relevance to Holland Bloorview: 

This research underscores the heightened challenges faced by families of children with disabilities during COVID-19, informing the need for specialized interventions. These findings can be utilized to develop targeted support programs, ensuring effective assistance for families navigating these difficulties.

Bio:

Jocelyn Aryeetey is a 4th year student at the University of Toronto, pursuing a double major in Biology for Health Science and Sociology. She is passionate about bridging gaps in healthcare and support services for children with disabilities. Under the guidance of Dr. Lamptey, Jocelyn is researching parental concerns on patience and behaviour management between parents with and without children with disabilities during the COVID-19 pandemic. Her work aims to highlight disparities in support systems during global crises. Jocelyn aspires to craft impactful healthcare solutions that contribute meaningfully to the advancement of health equity and inclusion.

University and Program: McMaster University - Mechanical Engineering

Supervisor and Lab: Dr. Jan Andrysek - PROPEL Lab

Project Title: Utilizing digital technology to develop a workflow for creating personalized and accessible prosthetic covers

Principal Investigator: Dr. Jan Andrysek

Research Summary:

Prosthetic covers are outer layers or shells designed to fit over prosthetic limbs and allow amputees to personalize the look of their device. Utilizing digital technologies such as computer-aided design (CAD) can allow clinicians to create prosthetic covers in-house at a lower cost, increasing the accessibility of these cosmetic components.

Research Question: 

Develop an efficient and cost-effective workflow for designing lower limb prosthetic covers using digital technology

Methods & Analyses: 

Using CAD software, parametric designs were created to adjust the cover based on a 3D scan of a patient’s lower limb features, such as ankle shape, and calf height and shape. Digital models were iteratively refined to optimize fit and alignment. 3D printing was employed to produce prototypes for initial evaluation, and clinician feedback guided further refinements.

Results: 

Clinician feedback led to key design improvements, such as more secure connection methods and adjustments to accommodate a larger range of prosthetic components.

Conclusions: 

A simple, adjustable digital workflow for designing lower limb prosthetic covers can provide an accessible option for personalizing and improving satisfaction of a prostheses. Next steps include adjusting the design for a wider range of prosthetic knees and conducting further patient testing.

Relevance to Holland Bloorview: 

By bringing this process in-house, the Holland Bloorview OrthoƟcs & ProstheƟcs department can offer enhanced funcƟonal and aestheƟc customizaƟons for a larger proporƟon of the client populaƟon.

Bio: 

Karthik is a second-year student going into third year in Mechanical Engineering at McMaster University. He is currently working in the PROPEL Lab, focusing on developing a cost-effective design and workflow for lower-limb prosthetic covers. His goal is to create a user-friendly workflow using computer-aided design while ensuring that the design accommodates the diverse needs of clinicians and clients for various lower limb prosthetics. Karthik is grateful for the opportunities and support provided by the PROPEL Lab, which have deepened his understanding of the mechanical design process.

University and Program: University of Waterloo - Systems Design Engineering

Supervisor and Lab: Dr Tom Chau, PRISM Lab

Project Title: Hierarchical combination of unsupervised clustering and linear classification methods to enhance personalized brain-computer interface classifiers.

Principal Investigator: Dr Tom Chau

Research Summary:

Brain computer interfaces (BCIs) allow users to control devices by performing mental tasks and modulating their brain signals. However, BCI use is limited by poor performance for some users. Analysis suggests the brain signals of these users are dominated by task-irrelevant neural activity. Consequently, task-related modulations are occluded and algorithmic ability to recognize task-specific patterns is impeded. This suggests that employing methods to adapt to these task-irrelevant modulations could improve classifier performance

Research Question: 

Does combining unsupervised clustering with traditional linear classifiers for BCI task decoding improve prediction accuracy?

Methods & Analyses: 

Data was collected from 14 adolescents. K-means clustering was applied to segmented each participant's data into regions of similar neural patterns. For lower performers, these regions represented different task-irrelevant patterns. Within each region, classifiers were deployed to make task predictions. We hypothesized that making predictions within these sub-regions would improve the discernibility of the data based on task-related modulations. To evaluate the approach, we compared the classification F1 scores to those obtained with a common BCI classifier.

Results: 

Averaged F1 scores across participants showed that the proposed scheme modestly outperformed the baseline classifier (0.64±0.15 vs. 0.56±0.14). Subsequent analysis showed that the proposed method achieved superior results in all but one participant.

Conclusions: 

This analysis yields insight into the task performance characteristics of low-performing users. These insights could guide personalized BCI design practices that would improve their clinical feasibility

Relevance to Holland Bloorview: 

BCIs could provide children with neurodevelopmental disabilities greater access to environmental interaction. This research informs how BCI design could be improved to meet the performance capabilities of individual users and increase accessibility to these technologies.

Bio: 

Maddie is a second-year student at the University of Waterloo, pursuing a BASc in Systems Design Engineering. She is passionate about harnessing the intersection of healthcare and technology to make meaningful impacts to society. This summer, Maddie worked alongside PhD candidate Nicolas Ivanov in Dr. Tom Chau’s PRISM lab to enhance personalized brain-computer interface (BCI) classifiers. With a goal to make BCIs accessible to a wider demographic, Maddie developed a hierarchical machine learning algorithm to help meet each individual's unique performance capabilities. Additionally, Maddie contributed to an open-source software library that assists with creating pipelines for modular BCI data processing.

University and Program: McMaster University - Health Sciences - Child Health

Supervisor and Lab: Dr Tim Ross, EPIC Lab

Project Title: Supportive Housing Experiences of Residents with Developmental Disabilities

Principal Investigator: Dr Tim Ross

Research Summary:

The need for supportive housing for PWDD has reached crisis levels, with thousands of PWDD spending years on waitlists. The short supply of supportive housing has led to some PWDD living with parents late into adulthood, or being placed in long-term care, which may not satisfy their needs. Scholars have given little attention to the supportive housing experiences of PWDD. This project aims to explore the experiences of PWDD within supportive housing residences, while investigating areas for improvement.

Research Question: 

How do PWDD experience their supportive housing environments?

Methods & Analyses:

Supportive housing residents participated in interviews and arts activities. Data were transcribed and thematically coded.   

Results: 

Preliminary analysis suggests staff support and design changes are needed to strengthen resident independence and wellbeing. Residents desire improved programming to bolster social opportunity. Actions are needed to ensure all residents feel a sense of belonging within the residence community.

Conclusions: 

The design, operations, and programming options within supportive housing residences require improvement to enhance residents’ independence and social belonging. Inclusion of resident voices in the planning, design, and operation of supportive housing could enhance residents’ supportive housing experiences and wellbeing. 

Relevance to Holland Bloorview: 

Study findings will aid Holland Bloorview (HB) clients with preparing for transitions into adulthood, and by enhancing their future housing options.  This research aligns with HB’s commitment to co-creation and IDEAA, as it involves and empowers the voices of PWDD to help inform accessible and inclusive supportive housing communities.

Bio: 

Ramneek is a third-year student at McMaster University, pursuing a Bachelor of Health Sciences, with a specialization in Child Health. She is passionate about learning more about childhood disability, and ways to optimize social systems to improve accessibility and inclusivity. As a Ward Summer Student, Ramneek works alongside Dr. Tim Ross to explore the experiences of people with developmental disabilities in supportive housing. Through her research, she is investigating ways to enhance the planning, design, and operations of supportive housing to promote the creation of accessible and inclusive communities.

 

University and Program: University of Toronto – Neuroscience, Computer Science

Supervisor and Lab: Dr. Azadeh Kushki, Autism Research Centre

Project Title: Implementing an AI model to predict the medication class prescribed to neurodivergent children

Principal Investigator: Dr. Azadeh Kushki

Research Summary: 

There are no biological markers to guide the selection of psychotropic medications for neurodivergent children. Artificial intelligence (AI) has the potential to fill this by providing data-driven insights to augment and inform clinical decision making.

Research Question: 

Is it possible to predict the medication prescription for neurodivergent children at a greater-than-chance accuracy using artificial intelligence?

Methods & Analyses: 

Two datasets were used to develop and evaluate an AI model for medication prediction: (1) the Holland Bloorview Psychopharmacology Program (PPP; n=277; age=10.50 (2.88) years, 83.75% male (sex), 26% intellectual disability), and (2) Province of Ontario Neurodevelopmental Disorders network (POND; n=499; age=10.82 (3.20) years, 68.73% male (sex), 73.34% white (race), 8.82% intellectual disability). Random forest was employed to generate medication recommendations for stimulant, antipsychotic, and antidepressant drug classes. For PPP, we trained our model on the first clinic visit and predicted recommendations for a follow-up visit. For POND, we performed a stratified 10-fold cross-validation. Receiver operating characteristic area under the curve (ROC-AUC) scores were used for model evaluation.

Results: 

Preliminary analysis shows good to excellent ROC-AUC scores across all drug classes (AUC>0.82 for PPP; AUC>0.76 for POND).

Conclusions: 

Our results demonstrate the potential of deploying personalized AI-based medication recommendation systems for neurodivergent children.

Relevance to Holland Bloorview: 

AI-based clinical decision aid holds the potential to inform decision making and provide personalized and timely care to neurodivergent children

Bio:

Ricky is a third-year student at the University of Toronto, pursuing a double major in Neuroscience and Computer Science. He is deeply interested in exploring the intersection of biology and computer science, particularly pertaining to research in medicine and disabilities. In his first summer as a Ward Summer Student, Ricky has been working alongside Masters Student Harshit Bokadia, led by Dr. Azadeh Kushki, to develop an AI model with the capability to predict the medication class of prescriptions given to neurodivergent children. He deeply appreciates the immersive learning opportunities presented to him by the Autism Research Centre and the Ward Family Summer Student Research Program.

University and Program: York University - Honours Biomedical Sciences, Minor in Psychology

Supervisor and Lab: Dr. Danielle Baribeau, Facilitating Individualized Interventions in Psychiatry and Neurodevelopment (FIIND) Lab under the Autism Research Centre

Project Title: Exploring the Association Between Financial Barriers to Care and Caregiver Burnout in Families of Children with Neurodevelopmental Disorders

Principal Investigator: Dr. Danielle Baribeau

Research Summary:

Caregivers of children with neurodevelopmental disorders (NDDs) face a high risk of burnout, exacerbated by financial barriers to accessing care. This study investigates the prevalence of financial barriers and their association with caregiver burnout in families of children aged 6-18 with NDDs receiving complex behavioural care at Holland Bloorview. By reviewing records from 343 children between 2019 and 2022, financial barriers were identified through clinician documentation and social services received, while caregiver burnout was documented in medical records.

Research Question: 

What is the prevalence of financial barriers to accessing care, and how is this associated with the experience of caregiver burnout in families?

Methods & Analyses: 

Conduct a Chi Square contingency test to find an association, as well as a logistical regression.

Results: 

Preliminary analyses revealed a significant association between financial barriers and caregiver burnout (X 2=42.909; p<.0001). Logistic regression models showed that financial barriers were a significant predictor of burnout, and not other confounding variables

Conclusions: 

Many families of children with NDDs face financial barriers to care, closely associated with caregiver burnout. Addressing these financial challenges may help in alleviating caregiver strain.

Relevance to Holland Bloorview: 

Comprehensive care should include support for caregivers’ mental health, respite access, and financial resource navigation to reduce caregiver strain. Systems and resources are needed to ensure equitable access to therapies and services for children with NDDs.

Bio: 

Sajeela is entering her third year at York University, studying Biomedical Sciences, minoring in Psychology. She had the pleasure of working with Dr. Baribeu and her team in this summer's Ward Summer Student program. Throughout her time here, she conducted a systematic chart review of patients from Holland Bloorview’s psychopharmacology clinic which allowed her to have a deeper grasp on the holistic aspects of client and family care. This inspired her research question examining the relationship between financial barriers to care, and caregiver burnout. She hopes that, in the near future, comprehensive care can consist of support for caregivers’ mental health, respite access, and financial resource navigation to reduce caregiver strain.

University and Program: University of Toronto, Saint George- Honours Bachelors of Arts-Social Sciences, Critical Studies in Equity and Solidarity, Anthropology, and Women and Gender Studies

Supervisor and Lab: Dr. Sally Lindsay, TRAIL Lab

Project Title: Understanding the experiences and disparities of sex/gender minoritized-youth with disabilities: A scoping review.

Principal Investigator: Dr. Sally Lindsay

Research Summary:

Sex/gender-minoritized (SGM) youth with disabilities, relative to non-disabled SGM youth, often experience heightened discrimination and exclusion, attributed to their intersecting marginalized identities. The intersection of gender identity, sex, and disability manifests into a surplus of economic, social, personal, and physical barriers. SGM youth with disabilities have been identified to experience exacerbated stigmatization, invalidation, and delegitimization following exploration and expression of sex/gender identity. Additionally, recurring social stigma and family-centered misinterpretations of gender identity and disability increase youths' probability of experiencing greater mental and physical health disparities.

Research Question: 

 What are the experiences and disparities of sex/gender-minoritized (SGM)  youth and young adults with disabilities? 

Methods & Analyses: 

A scoping review was conducted, drawing on 6 databases, where two researchers screened titles and abstracts. 23 studies met our inclusion criteria. A narrative-synthesis approach was utilized to identify key trends.

Results: 

Identified main themes associated with SGM youth with disabilities: (1) exploration and expression of sex/gender identity (i.e., concealment, negotiation, disclosure, invalidation); (2) adverse experiences (i.e., harassment, discrimination, abuse); (3) impact on mental health and quality of life; and (4) coping techniques (i.e., family, resources, supports, safe environment).

Conclusions: 

Several disparities and barriers exist for SGM youth with disabilities that affect their well-being and ability to affirm gender identity. Additional research is required to explore the prolonged effects of these negative experiences and impacts.

Relevance to Holland Bloorview: 

Focusing on the experiences and impact of both SGM and disabled identities within youth, including family concerns, enables staff to adequately address and situate the lived experiences, challenges, varying needs, and support systems of Holland Bloorview's diverse clientele. Collecting SGM-disabled based experiences from clients and families should be considered to implement possible intervention strategies and supports.

Bio:

Teya is entering her fourth year at the University of Toronto, pursuing a Honours BA in Social Sciences with a major in Critical Studies in Equity and Solidarity, and minors in Anthropology and Women and Gender Studies. She is immensely passionate about disability inclusion, awareness, and education, exemplified by her involvement as a student leader in three UofT accessibility-based organizations, the creator of the UofT accessibility podcast, and as a volunteer at Holland Bloorview. Teya is a member of the TRAIL Lab, run by Dr. Sally Lindsay, and here she has applied her interdisciplinary skills of knowledge translation to analyze how youth with co-occurring marginalized identities endure specific disparities that jeopardize their well-being. Lastly, she is leaving the WARD program with an enhanced awareness of the importance of intersectionality in disability research and an appreciation for all that Holland Bloorview offers.

University and Program: York University - Honours Biomedical Science

Supervisor and Lab: Dr. Evdokia Anagnostou and Dr. Jacob Ellegood, Autism Research Centre (ARC)

Project Title: Whole brain analysis to examine the relationship between brain measures and language functionality in children with neurodevelopmental disorders

Principal Investigator: Dr. Evdokia Anagnostou

Research Summary: 

Most research on language ability and the brain has been done using Clinical Evaluation of Language Fundamentals (CELF). There are currently no studies on this interaction using the OWS-II, measured using the Oral and Written Language Scales, 2nd ed. (OWLS-II) despite it being a widespread, validated measure of English language proficiency. In addition, multimodal, large scale and transdiagnostic research in oral language functionality is limited. In our research, we used the OWLS-II to measure listening comprehension and oral expression as well as overall language ability in relation to brain physiology and activity through a combination of neuroimaging techniques.

Research Question: 

Can the OWLS-II be applied to determine the relationship between oral language and brain measures in neurodivergent (ND) children [(Attention deficit hyperactivity disorder (ADHD) and Autism spectrum disorder (ASD)] compared to typically developing (TD) children?

Methods & Analyses: 

We conducted a cross sectional analysis of data obtained from a convenience sample of ND - ADHD and ASD - and TD children (3-21 years). Participants were administered the OWLS-II by a clinician to assess for oral language functioning and completed structural and functional magnetic resonance imaging (sMRI and fMRI) scans. Correlational analyses were conducted using ANOVAs to explore the relationship between OWLS-II and various regions of the brain.

Results: 

Of 2,027 participants, 493 had ADHD, 1,062 had ASD and 472 were TD. The mean age was 11 years. Strong correlations were observed between OWLS-II scores and over 14 regions across the brain (r<0.05) by diagnosis, including the left calcarine fissure, right lingual gyrus, left middle occipital gyrus, left cuneus, right gyrus rectus, right insula, right supramarginal gyrus and right inferior frontal gyrus. Particularly, the left middle occipital gyrus was incidentally found to be inversely correlated with speaking ability in typically developing children and positively correlated in those with ASD and ADHD.

Conclusions: 

This study supports the effectiveness of the OWLS-II for measuring language functionality, demonstrating its accuracy as similar to the CELF in clinical research. It also suggests that some brain regions may have a different influence on language functionality in ND and TD children. It identifies possible regions for further research based on differences between diagnoses.

Future Directions:

Further investigation will incorporate other neuroimaging techniques, such as diffusion tensor imaging (DTI) and electroencephalography (EEG) in addition to sMRI and fMRI to conduct a comprehensive multimodal whole brain analysis to further examine brain measures and their relationship with language ability in ND populations.

Bio:

Vivien is entering her third year in Biomedical Science at York University. She has enjoyed learning about the impactful work done through the Ontario Brain Institute (OBI), Province of Ontario Neurodevelopmental Disorders Network (POND) and Holland Bloorview Rehabilitation Hospital. Her research focuses on the relationship between brain activity and connectivity and language impairments across neurodevelopmental disorders. During her placement, she has appreciated exchanging ideas with other students and learning more about transdiagnostic research as well as its applications in neurobiology. She hopes to contribute to expanding field knowledge to improve diagnoses with neuroimaging and providing children with neurodevelopmental differences more effective care.

University and Program: York University, Specialized Honours Psychology

Supervisor and Lab: Dr. Melanie Penner, Autism Community Capacity and Evaluation of Programs and Training(ACCEPT) Lab under Autism Research Centre (ARC)

Project Title: Assessing facilitators and barriers to pediatric autism assessment using an implementation science framework

Principal Investigator: Dr. Melanie Penner

Research Summary: 

Autism is a neurodevelopmental condition impacting various aspects of one’s life. Early diagnosis is key to gaining timely support to impact skill acquisition. Though some studies explore frustrations related to autism care, the current study uses an implementation science framework to better understand barriers and facilitators pediatricians face with diagnosis, facilitating identification of corresponding strategies. The current research is based on Ontario data from a multi-province study.

Research Question: 

What are current diagnostic facilitators and barriers community pediatricians across Ontario face in conducting autism assessments? 

Methods & Analyses: 

Invitation letters were sent to ECHO Ontario Autism participants who met the inclusion criteria. Seven community pediatricians participated in a focus group or individual interview facilitated by the research manager or principal investigator. The interview guide followed Theoretical Domains Framework (TDF) guidelines, and transcripts were coded and analyzed by a summer research student and research coordinator using reflexive thematic analysis to identify relevant themes. These themes were then categorized deductively using the TDF through NVivo.

Results: 

Preliminary analyses highlight main themes of facilitators and barriers to conducting autism assessments. Facilitators include experience, knowledge around resources, and supportive networks, while barriers include assessing complex cases, a lack of targeted training towards assessment, and inflexible policies.

Conclusions: 

Although preliminary results highlight diagnostic facilitators and barriers in Ontario, additional provincial interviews will help recognize across-site and site-specific needs.

Relevance to Holland Bloorview: 

Recognizing facilitators and barriers influences policy makers and clinicians to implement necessary changes, improving early diagnosis and resource access for families.

Bio:

Zahra is a third-year student at York University, pursuing a Specialized Honours Psychology degree. She is deeply interested in accessible care with a focus on mental health, and works to recognize and alleviate inequities in healthcare. Much of her work is around children’s health, development, resilience, and social inclusion, with a focus on children with developmental delays or differences. She’s grateful for the opportunity to work with Dr. Melanie Penner and her team at the ACCEPT Lab. Her project focuses on identifying facilitators and barriers pediatricians face when conducting autism assessments to help develop toolkits for clinicians, encouraging more assessments and improving diagnosis and care access for families. 

University and Program: McMaster University - Integrated Biomedical Engineering and Health Sciences

Supervisor and Lab: Dr. Azadeh Kushki, Neurodiversity and Personalized Health Lab under the Autism Research Centre

Project Title: Characterizing sociodemographic biases in adaptive functioning data in a cohort of neurodiverse children

Principal Investigator: Dr. Azadeh Kushki

Research Summary: 

Adaptive functioning quantifies the ability to independently perform skills needed in everyday life. Beyond clinical causes, adaptive functioning may be impacted by disparities in access to timely services due to sociodemographic factors.

Research Question: 

To what extent do sociodemographic factors bias child adaptive functioning scores?

Methods & Analyses: 

Social, conceptual, practical, and general adaptive composite (GAC) scores from the Adaptive Behaviour Assessment System questionnaire were obtained from the Province of Ontario Neurodevelopmental Disorders network. The dataset included 1254 neurodiverse participants (mean age = 10.29±4.33 years; male = 872; White ethnicity = 1064; above median income = 805). Higher scores indicate better adaptive functioning. Differences in composite scores across subgroups defined by sex (male/female), ethnicity (White/non-White), income level (greater/less than provincial median), primary caregiver’s education (having/lacking post-secondary education), and caregiving environment (equal/single caregiver) were examined using Mann-Whitney U-tests. An XGBoost machine learning model was employed to predict composite scores, categorized as above/below 70, using sociodemographic features with stratified 15-fold cross-validation. Receiver operating characteristic area under the curve (AUC) scores were used for model evaluation.

Results: 

Higher GAC scores were associated with female sex (p < .001), above-median income (p < .001), caregiver education (p < .001), caregiving environment (p < .001) and White ethnicity (p = .01). The median AUC across folds was .688±.128. Household income and sex were the most important predictors of all composite scores.

Conclusions: 

Our results demonstrate that adaptive functioning significantly differs between subgroups based on socioeconomic status, sex, and ethnicity.

Relevance to Holland Bloorview: 

This research provides precision health insights into sociodemographic predictors of adaptive functioning, informing personalized approaches to supporting neurodivergent children and their families according to Holland Bloorview’s Transformative Care, Inclusive World plan

Bio:

Zuhair is entering his second year at McMaster University in the Integrated Biomedical Engineering and Health Sciences program, majoring in Software and Biomedical Engineering. He is interested in further exploring the applications of software and data science to health-related research and eventual therapeutic development, especially in care for vulnerable populations. This summer, Zuhair worked closely alongside his supervisor, Dr. Azadeh Kushki, and research engineer Harshit Bokadia in the Autism Research Centre at Holland Bloorview, using machine learning and statistical testing to identify and study sociodemographic biases and clinical differences in adaptive functioning data of neurodiverse children. Through his work, Zuhair has developed a deeper understanding of data preprocessing, clinical applications of machine learning, and measurements of fairness in predictive models. He is grateful to have had the opportunity to partake in this unique project that has widened his understanding of the persisting inequities between demographics in Ontario’s healthcare system.