Platform Technology for Neurodegenerative Diseases Landscape and Roadmap

Platform Technology for Neurodegenerative Diseases: Landscape and Roadmap development Jude Bek Computer Science/ Neuroscience & Experimental Psychology University of Manchester

Objectives Centralising research initiatives in technology for ND - Across Uo. M faculties and schools/divisions • Strengthening funding bids & external collaborations • Working with Dementia@Manchester • Support from Manchester Informatics http: //bit. do/Neuro. D-PT-Launch 2

Connecting technological capability and clinical need Tech: Clinical: Objective/quantitative measurement Diagnosis Continuous/frequent data collection Heterogeneous data Algorithms/models Monitoring Treatment Solutions: Communication Biomarker identification Profiling/stratification Individualised treatment/Precision medicine Cost-effective interventions Living well http: //bit. do/Neuro. D-PT-Launch 3

Technological solutions for neurodegeneration Clinical assessments are typically: • • • Subjective – clinician/patient/carer Categorical Temporally constrained – clinic visits Unrepresentative – stress/ecological validity Insensitive to subtle change Outdated? Going beyond traditional clinical measures: • • Scope/complexity of behaviours Sensitivity to fluctuation/change http: //bit. do/Neuro. D-PT-Launch 4

Technological solutions for neurodegeneration • • • Capitalising on increasing tech use in older adults Provision of detailed data – tailored for clinician/patient Efficiency Cost effectiveness – increasingly stretched healthcare resources Convenience Psychological burden –confrontation with symptoms/disease progression (test 'failure'; reflecting on difficulties) http: //bit. do/Neuro. D-PT-Launch 5

Clinical snapshots: motor and non-motor symptoms http: //bit. do/Neuro. D-PT-Launch 6

Clinical snapshots: mood http: //bit. do/Neuro. D-PT-Launch 7

Methods • Mapping the landscape • • Catalogue current research in ND Identify synergies; cross-cutting themes • Developing a roadmap • • Sharing of data and resources Opportunities for collaboration New tools and new applications Ongoing communication/consultation http: //bit. do/Neuro. D-PT-Launch 8

Outputs • White paper: landscape/roadmap • Integrated research community • Web resource • Newsletter • Seminars • Workshops …continued partnerships http: //bit. do/Neuro. D-PT-Launch 9

Ce. HRes Roadmap http: //bit. do/Neuro. D-PT-Launch 10

Landscape development Surveying and cataloguing existing or potential applications for ND across schools Identifying: • Themes • Synergies/overlaps • Gaps • Opportunities • Challenges http: //bit. do/Neuro. D-PT-Launch 11

Landscape development • Bringing together: • Technologists with health researchers • Technologists with technologists • Technologists with clinicians • Identifying contributors: • 9 schools across FSE • FBMH http: //bit. do/Neuro. D-PT-Launch 12

Landscape development Reaching out to technologists across FSE • Contacting schools • ‘Pop-up’ information sessions • Follow-up (drop-in) sessions Information gathering • Workshops: 1) Health + Tech: 31 Jan -1 Feb, Shrigley Hall Hotel 2) Tech + Tech: 7 -8 March, MOSI 3) All – Roadmap development: TBC • Interviews http: //bit. do/Neuro. D-PT-Launch 13

Landscape development Pop-up sessions • • February 2017 In-school advertising Presentations, Q&A Informal follow-up • Identification of school representatives http: //bit. do/Neuro. D-PT-Launch 14

Landscape development (1) Current projects in ND • Background /objectives • Partners/collaborators • Development/methods • Data • Results/interpretation • Next steps http: //bit. do/Neuro. D-PT-Launch Interview topics Challenges • Technical/resources • Ethics/regulatory • Recruitment/compliance • Interpretation • Validation (2) Potential projects • Re-purposing • Technologies in development • Potential collaborators? • Input needed? 15

Connecting technological capability and clinical need Tech: Clinical: Objective/quantitative measurement Diagnosis Continuous/frequent data collection Heterogeneous data Algorithms/models Monitoring Treatment Solutions: Communication Biomarker identification Profiling/stratification Individualised treatment/Precision medicine Cost-effective interventions Living well http: //bit. do/Neuro. D-PT-Launch 16

Project examples Monitoring in Parkinson’s disease and dementia Collaborative projects (FSE/FBMH): • Symptom Knowledge in Parkinson’s (SKIP) • • • PI: Ellen Poliakoff (Manchester) Objective: monitor symptoms across multiple domains in Parkinson’s Software Architecture for Mental Health Self. Management (SAMS) • • PI: Peter Sawyer (Lancaster) Objective: detect early signs of cognitive decline http: //bit. do/Neuro. D-PT-Launch 17

Examples: Parkinson’s disease and Dementia Parkinson’s Dementia Constructs measured Cognition, speech, social, motor Cognition (memory, planning…); functional behaviour Feasibility/ Proof of concept Directed tasks (laptop) x 30 days – detection of fluctuations Video diary Focus group Semi-directed tasks –crosssectional; group differentiation (MCI/mild AD vs control) Expert consultation Continuous monitoring -passive data collection -proxy measures -validation against clinical Smartphone data: physical activity, communications, location, environment, web/ interactions Daily home PC use: mouse/keyboard, operating system, web/email, text analysis -e. g. Addenbrooke’s Cognitive -e. g. UPDRS Examination http: //bit. do/Neuro. D-PT-Launch 18

Parkinson’s disease http: //bit. do/Neuro. D-PT-Launch 19

Parkinson’s disease http: //bit. do/Neuro. D-PT-Launch 20

Dementia Data mining Layer II SAMS participant Raw data Layer III Layer IV http: //bit. do/Neuro. D-PT-Launch Raw text Computer-use behaviours Text mining Cognitive/ functional change? 21

Dementia Keystrokes/min 35 30 25 20 15 10 5 0 Keyboard behaviour HC MCI/AD Characters/min 150 Characters/min (Typing task) 250 200 150 r =. 598, n = 41, p <. 001* 100 50 100 0 0 50 0 HC http: //bit. do/Neuro. D-PT-Launch MCI/AD 10 20 30 40 50 Keystrokes/min (Total experiment) 60 70 r =. 592, n = 41, p <. 001* 22

Synergies and Challenges Proof of concept -Constructs to measure -Cross-sectional analysis -User/expert consultation Identification of suitable proxies Retention/ compliance Feasibility -Small-scale data collection -Patterns/fluctuations Recruitment Longitudinal data collection -Larger sample size -Change detection Acceptability 23 http: //bit. do/Neuro. D-PT-Launch

Synergies and Challenges Raw data -GPS -Web -Sensor -Bluetooth -Keyboard/mouse Hardware/ software limitations Large volumes of data http: //bit. do/Neuro. D-PT-Launch Analysis -Identifying metrics -Patterns -Variation Combining data sources Inferring complex behaviours Interpretation -Disease progression -Functional decline -Individual profiles Correlating with clinical measures Missing data 24

Overlaps and themes • Unobtrusive monitoring – everyday activity • Utilising off-the shelf devices/apps • Home PC use • Smartphones – GPS, bluetooth… • • Heterogeneous data Pattern identification and interpretation Change detection Validation/mapping to clinical measures http: //bit. do/Neuro. D-PT-Launch 25

Future opportunities and challenges Next steps: • • SKIP – 12 month longitudinal study planned SAMS – longitudinal data analysis under way Opportunities: • • Application to other conditions? Combining methods - PC use and sensors? – cognitive, physical and social domains Challenges: • • Continued collaboration Implementation pathway http: //bit. do/Neuro. D-PT-Launch 26

Importance of collaborative working • Appropriate/timely research questions • Combined expertise needed to analyse/interpret data • User/patient input needed • Feasibility • Acceptability • Consultation at every stage • E. g. SKIP focus group: -patients valued passive monitoring -level of feedback important http: //bit. do/Neuro. D-PT-Launch 27

Get in touch… http: //www. informatics. manchester. ac. uk/ neurodegenerative-diseases-platformtechnology/ http: //bit. do/Neuro. D-PT-Launch 28

Acknowledgements Sam Couth Julio Vega Gemma Stringer SAMS and SKIP study teams

Questions? Jude Bek Judith. bek@manchester. ac. uk