SiLab

Welcome to the Stroke Innovation (& Machine learning) Lab [SiLab]

Transforming Neurovascular & Stroke Care Through Machine Learning Intelligence: #AI4QI; Our lab focuses on Quality Improvement in stroke and neurologic care. Our research domain is clinical neurology, with a particular emphasis on neurovascular care. In my quality improvement (QI) lab, we concentrate on stroke care and utilize machine learning techniques applied to datasets and time-series data. Recently, our focus has shifted to exploring sound as a biomarker. I am the Principal Investigator, Dr. Khosravani, from the Division of Neurology at the University of Toronto.

Areas of academic focus:

• #AI4QI or ML4QI - We pioneer machine learning applications to elevate quality improvement in acute stroke care. Our flagship initiative, Project MASA (Machine Learning Assisted Swallowing Assessment), analyzes voice biomarkers to revolutionize swallowing assessments and enhance patient outcomes.

• Stroke Resuscitation Innovation - As originators of the “stroke resuscitation” concept and Protected Code Stroke protocol, we lead the integration of Crisis Resource Management (CRM) principles into acute stroke care. Our simulation program runs across the emergency department, code stroke nursing, neurology residency, the neurovascular unit, and the angiography suite, and advances team performance and patient safety during critical moments of stroke intervention, complementing our broader work with the Neurovascular Innovations CollaborativE (NICE). See the full project page and our simulation resources.

• Integrative Stroke Palliative Care - We champion comprehensive stroke care by developing evidence-based protocols that seamlessly integrate palliative medicine into acute stroke treatment. Our research defines best practices for incorporating palliative approaches while maintaining aggressive intervention when appropriate, ensuring optimal patient-centered outcomes across the care spectrum.

Members of the lab are also involved in educational initiatives, including a podcast on stroke education: Stroke FM Podcast, the official podcast of the Canadian Stroke Consortium.

Stroke Innovation-machine learning Lab (SiLab) machine learning for quality improvement

• My lab is part of the NQIL group, at the University of Toronto; a hub for QI for neurology.
• We also are exploring the intersection of machine learning and quality improvement, utilizing voice-based technologies to refine our methods. We have expertise in deep-learning and processing of audio signals.

• Our pursuit of excellence extends to the cutting-edge field of machine learning. With support from T-CAIREM and SHSC, we are leveraging bedside physiologic recordings to improve the quality of acute stroke care.

• Machine learning as applied to quality improvement in stroke
  • 2026-2027 (academic years)
    • Lab Members:
      • Dr. Emmanuel Adegunna, Neurology, PGME Neurology, University of Toronto, PGY5 - palliative care in stroke; joined project Pal-MASA on July 1, 2026 as his PGY5 resident research project
      • Dr. Bension Tilley, Neurology, PGME Neurology, University of Toronto - working on the use LLMs in EHRs, and project 1stroke.org
      • Dr. Arjun Balachandar, MD, FRCPC, Neurology, University of Toronto - independent Principal Investigator and collaborator of the lab (lab alumnus), Movement Disorders Fellow completing his PhD in the Clinician-Scientist Training Program, part of project MASA
      • Dr. Rami Saab, Neurology Resident, University of Toronto (Temerty Faculty of Medicine graduate) - ML4QI in Stroke, part of project MASA
      • Rebecca Wu, UGME, U of T Medicine (Temerty Faculty of Medicine, Class of 2027), working on LLMs in Neurology
      • Luke Zhang, Computer Science and Statistics, U of T, T-CAIREM (2026 student award) - CNN/ViT-based dysphagia screening from continuous speech, part of project Pal-MASA
      • Richard Zhang, MD Program, McMaster University
      • Dr. Hamza Mahdi, MD (Schulich School of Medicine, Western University), Orthopedic Surgery, University of Toronto - ML4QI in Stroke, part of project MASA
  • 2025-2026 (academic years)
    • Lab Members:
      • Rishit Dagli, Computer Science, U of T, T-CAIREM (2024 student award), now at NVIDIA doing a 1-year studentship, worked on ML4QI in Stroke in project MASA (machine learing assisted swallowing assessment), and pioneered and developed SEE2SOUND
      • Dr. A. Balachandar, Neurology, U of T, Movement disorders Fellow and pursuing PhD in movement disorders and DBS, worked on ML4QI in Stroke in project MASA (machine learing assisted swallowing assessment)
      • R. Saab, U of T Med, ML4QI in Stroke, works on ML4QI in Stroke in project MASA (machine learing assisted swallowing assessment)Sunnybrook Research Institute
      • E. Nashnoush, MSc, U of T Data Science, ML4QI in Stroke, Sunnybrook Research Institute, T-CAIREM, HealthQuality, part of Project MASA
      • H. Mahdi, Western University, Medicine, ML4QI in Stroke, part of project MASA
      • Dr. B. Tilley, Neurology, PGME Neurology, University of Toronto - working on the use LLMs in EHRs, and project 1stroke.org
      • Dr. E. Adegunna, Neurology, PGME Neurology, University of Toronto - working on palliative care in stroke
      • Diya Ahmed, UGME, U of T Medicine, working on QI in Neurology and future project on LLMs in Neurology
      • Rebecca Wu, UGME, U of T Medicine, working on LLMs in Neurology
      • Jake Lance, UG, U of T
  • 2024-2025 (academic years)
    • Lab Members:
      • Rishit Dagli, Computer Science, U of T, T-CAIREM (2024 student award), now at NVIDIA doing a 1-year studentship, worked on ML4QI in Stroke in project MASA (machine learing assisted swallowing assessment), and pioneered and developed SEE2SOUND
      • Dr. A. Balachandar, Neurology, U of T, Movement disorders Fellow and pursuing PhD in movement disorders and DBS, worked on ML4QI in Stroke in project MASA (machine learing assisted swallowing assessment)
      • R. Saab, U of T Med, ML4QI in Stroke, works on ML4QI in Stroke in project MASA (machine learing assisted swallowing assessment)Sunnybrook Research Institute
      • E. Nashnoush, MSc, U of T Data Science, ML4QI in Stroke, Sunnybrook Research Institute, T-CAIREM, HealthQuality, part of Project MASA
      • H. Mahdi, Western University, Medicine, ML4QI in Stroke, part of project MASA
      • Dr. B. Tilley, Neurology, PGME Neurology, University of Toronto - working on the use LLMs in EHRs, and project 1stroke.org
      • Dr. E. Adegunna, Neurology, PGME Neurology, University of Toronto - working on palliative care in stroke
      • Diya Ahmed, UGME, U of T Medicine, working on QI in Neurology and future project on LLMs in Neurology
      • Rebecca Wu, UGME, U of T Medicine, working on LLMs in Neurology
      • Sofia Berrada, UG, U of T, 2024 volunteer
      • Jake Lance, UG, U of T
  • 2023-2024
    • Lab Members:
      • Dr. A. Balachandar, Neurology, U of T, PGY4/5 Project, ML4QI in Stroke
      • R. Saab, U of T Med, ML4QI in Stroke, Sunnybrook Research Institute
      • E. Nashnoush, MSc, U of T Data Science, ML4QI in Stroke, Sunnybrook Research Institute, T-CAIREM (2023 student award)
      • H. Mahdi, Western University, Med, ML4QI in Stroke, Sunnybrook Research Institute
      • Dr. B. Sivanandan, Neurology, U of T, PGY4/5 Project, Palliative Care in Stroke
      • Dr. M. Mahendiran, U of T, Family Medicine (graduated), Palliative Care research, Hospital Medicine Fellow (Orange Team)
      • Georges Khalaf, CREMS, Summer 2024, UGME, U of T Medicine
      • Dr. E. Adegunna, Neurology, U of T, PGY2
      • Dr. B. Tilley, Neurology, U of T, PGY1
      • Diya Ahmed, UGME, U of T Medicine
  • 2022-2023
    • Lab Members:
      • Dr. A. Balachandar, Neurology, U of T
      • R. Saab, U of T Med, T-CAIREM (2022 student award)
      • M. Panchal, U of T Med, CREMS

Simulation in Code Stroke - Neurovascular Resuscitation - Neurovascular Innovations CollaborativE (NICE)

• Our lab championed the framework of crisis resource management in stroke simulation to optimize critical intervention metrics such as “door-to-needle” times. We are proud to be pioneers in the field and published the first reframing of Crisis Resource Management (CRM) for stroke care. We also developed the “protected code stroke” during the COVID-19 pandemic, which was integrated into national and international guidelines and downloaded 47,945 times from the American Heart Association and Stroke journal’s website, with 171 citations (publisher metrics, August 2026). Our research aims to enhance care pathways and human performance factors for acute stroke patients through the simulation of neurovascular resuscitation. We have introduced the concept of “neurovascular resuscitation,” applying principles from medical and trauma resuscitation to stroke treatment, thereby reinstating the ‘code’ in code stroke.
• In 2023, Dr. Khosravani and Dr. Hawkes co-founded the Neurovascular Innovations CollaborativE (NICE) @neuroresuslab, an initiative projected to contribute substantially to augmenting neurovascular care education. Our Lab’s focus at SiLab is the medical/neurocritical care aspects of hyperacute care and techniques.
• Through NICE, Dr. Khosravani and Dr. Hawkes began collaborating to expand the simulation work to include neurovascular EVT simulation, now a component of SRC-TO, SRC-FU, and SRC-Ottawa - a collaboration highlighted throughout our Provincial Stroke Rounds presentation.

The simulation program

• A code stroke is a resuscitation - time-critical, cognitively loaded, and delivered by a team that has often never worked together before. Unlike cardiac arrest or trauma, it has historically not been rehearsed as one. What began in 2019 as in situ simulation with our Emergency Department colleagues has grown into a multi-stream, interprofessional program:
  • Emergency Department simulation (from 2019) - the origin of the program, run in situ with ED physicians and nurses, focused on the first minutes of the code stroke.
  • Code Stroke RN (CSRN) simulation - interprofessional simulation with the code stroke nursing group, who are the connective tissue of every activation, from the EMS phone call onward.
  • Stroke Boot Camp - code stroke simulation for incoming PGY1 residents (2020-2025) - an annual July day at Sunnybrook for incoming neurology residents across the city-wide University of Toronto Division of Neurology. Dr. Khosravani started this orientation and pioneered running it before residency begins rather than after; over six years it grew from a simulation in a conference room into full sessions at the Sunnybrook Canadian Simulation Centre. As the orientation broadened to take in general neurology, he continued to lead its stroke resuscitation work - case selection and case scenario development - with our code stroke nurses joining when able.
  • SRC-TO - the Stroke Resuscitation Course, Toronto (April 2024), Canada’s first local Stroke Resuscitation Course.
  • SRC-FU - the follow-up course (January 2025), with a curriculum for neurology residents covering point-of-care ultrasound (POCUS) in hyperacute stroke assessment, a simulation centre phase, and EVT complications and considerations.
  • SRC-Ottawa - the national course (June 2025) - Canada’s first Stroke Resuscitation Course, run as a course at the Canadian Neurological Sciences Federation Annual Meeting with Dr. Khosravani as Course Director; both the first extension of the course beyond Toronto and its first national delivery. SRC-TO, SRC-FU and SRC-Ottawa were all funded through the Slamen-Fast New Initiatives in Neurology program; our thanks to the program and its donors.
  • B4 Neurovascular Unit skills days - nursing professional development days bringing resuscitation skills onto the inpatient neurovascular unit.
  • EVT sedation simulation (Sunnybrook) - a distinct, anesthesia-facing stream run in house with our anesthesia and endovascular colleagues, rehearsing sedation and airway decisions during thrombectomy; the first joint physician-nursing simulation for endovascular stroke treatment in Canada.
  • Door-to-CT (DT-CT) - a quality improvement stream now led by the code stroke nurses, translating what the simulations surface into measured process change.

Teaching, spread, and quality improvement

• Simulation here is not an end in itself; it is a quality improvement instrument. Alongside PDSA-cycle process redesign, median door-to-needle time at Sunnybrook fell 25%, from 38.5 to 29 minutes - below the national average and durable over eleven months - across a service managing more than 1,000 code strokes a year.
• The teaching infrastructure underneath the program is codestroke.net, running since 2017, with over 25,000 page-user instances across more than 9,100 unique users, and in use by trainees, stroke clinicians, and Ontario EMS.
• The work was presented at Provincial Stroke Rounds in May 2026 by Dr. Khosravani and Dr. Christine Hawkes, as Stroke Resuscitation Simulation: Building High-Performing Teams Through Crisis Resource Management, and the materials were subsequently requested for provincial distribution. Regional and inter-hospital collaborations, including transfer and EMS simulation with partner sites, are in active development.
• The program is run with the Sunnybrook Canadian Simulation Centre, together with colleagues in emergency medicine, code stroke nursing, the B4 neurovascular unit, anesthesia and endovascular therapy, and the Toronto Stroke Networks. Full details are on the stroke resuscitation project page.

Routine Integration of Palliative Care in Stroke

• Despite significant advancements in stroke care, a considerable number of patients still grapple with substantial morbidity and mortality. Recognizing this, we advocate for the routine integration of palliative care into stroke treatment. Compassionate and effective care forms the bedrock of the philosophy we advocate for in terms of expanding the confluence of palliative medicine and stroke care.

Prospective Students

• Each year my lab takes 1 student from T-CAIREM, and another student either via SRI (or CREMS).

• If you are a student with an interest in QI and experienced in research, or if you are an engineering, CS, or MD student interested in clinical applications of machine learning in neurology, please visit our Contact the Lab page to get in touch.

site disclaimer and description

• Please note: All content on this site, including blogs, links, topics, and posts, is intended solely for educational purposes. This site does not constitute medical advice or consultation and does not establish a duty of care or any medical care/information/advice/or other obligations. Opinions expressed here are personal and do not reflect or replace expert advice from any institution, workplace, or organization. By using the site you agree to this.