The SALSA system provides a fully-automated, quantitative solution for analyzing speech that can better evaluate and monitor patients with cognitive neurodegenerative conditions. It automates the three steps of recording a picture naming task, translating/recognizing the speech, and analyzing features of the speech. This quantitative method measures aspects like disease progression and response to therapy, addressing the need for objective tools. The system is compatible with existing clinical workflows and has potential market opportunities in pharma, hospitals, and for medical professionals.
1. System for Automated Language y g g
and Speech Analysis ‐ SALSA
Andrew MorrowAndrew Morrow
November 10, 2010
The data in this presentation is representative of theThe data in this presentation is representative of the
information known at the time of printing
(11/11/2010) and should be confirmed by the UMN
Technology Marketing Manager.
2. Office for Technology Commercialization 8
Why Automated Speech Analysis?
Cognitive neurodegenerative assessment batteries (e.g.
MTDDA, PICA & Boston) utilize Qualitative measurement
methods.
Patients worldwide: 75 Million
Growth: 5 Million per year
Alzheimer’s, Epilepsy, Aphasia
Quantitative methods are
needed to better evaluate
and monitor:
• initial disease diagnosis
• disease progression
• therapeutic response
3. Office for Technology Commercialization 9
Automated Speech Analysis: Solution
No commercially available system has introduced the
automation of all three steps of the assessment process:
1. Recording of picture naming task
2. Translation/Recognition
3. Analysis
The Automated Speech Analysis System is
a fully-automated quantitative solution
4. Office for Technology Commercialization 10
Automated Speech Analysis: Value
SALSA Method
Quantitative Assessment
Measures: Progression Monitoring, Response to Therapy, Post-Event Evaluation
6. Office for Technology Commercialization 12
Automated Speech Analysis:
Compatible with Existing Patient Flow
1. Diagnose using test 2. Confirm with EEG 3. Administer Therapy
4. Retest and compare to baseline
7. Office for Technology Commercialization 13
Automated Speech Analysis: Markets
Pharmaceutical Manufacturers
- Clinical validation of their new therapies
- Need to acquire quantifiable data
Hospitals
- Clinical validation of therapy changes and in-patient releases
- Now have sufficient reimbursement regulations
Medical Professionals
- Clinical validation to diagnose patients
- Eager to diagnose accurately
Key Step: Clinical Validation
8. Office for Technology Commercialization 14
Automated Speech Analysis: Report
Subject ID
Visit ID
Raw Scores
Scores normalized by sample duration
9. Office for Technology Commercialization 15
Automated Speech Analysis:
Business Model Options
Service
Recording uploaded to hosted server
Target: Clinicians
Shrink-wrap
Distribute software directly
Target: Pharmaceuticals
Hardware
Hand-held or other microphone integrated to the current
software to create an end-to-end testing platform.
Target: Hospitals
10. Office for Technology Commercialization 16
Automated Speech Analysis: Inventor
Serguei Pakhomov, Ph.D.
Assistant Professor
Department of Pharmaceutical
Care & Health Systems
Co-founder: UMN Center for
Clinical and Cognitive Neuropharmacology
Research Focus:
Current research areas include natural language and
speech processing and their applications in the
biomedical domain.
Related IP: Sets of picture naming stimuli for repeated
assessment of cognitive performance.
11. Office for Technology Commercialization 17
Automated Speech Analysis: Wrap-up
• New speech analytics algorithm
targeting Alzheimer’s and Epilepsy
patients.
• Patent Application published
• Compatible with existing testing
methodologies
• Seeking a commercialization partner