Cross-species Machine Learning-based Translation of Physiological Models to Improve Next-generation Clinical Peripheral Neurostimulation

Investigators
Caterina Stamoulis, Deniz Erdogmus, Theodoros Zanos, Min Zhang, Michael Miga
Contact info (email)
caterina.stamoulis@childrens.harvard.edu
1. Define context(s)
aid in clinical decision making
aid in clinical trial design
identify/explore new therapies
reveal new biological insights
Current Conformance Level / Target Conformance Level
Adequate
Primary goal of the model/tool/database

The goal of this resource is to develop a relatively generic framework that would allow preclinical experiments to inform clinical decision making.  Unfortunately in the area of peripheral neuromodulation, it is not possible to do analogous modulation as one can do in the preclinical environment.  The question this model-building framework is asking is, "Can human peripheral therapeutic stimulation be mapped to a preclinical system and produce an appropriate  electrophysiological response which is subsequently mapped to an appropriate human electrophysiological response all done in silico with deterministic and machine learning models?"  This framework would essentially functionalize the preclinical system as a forecaster/predictor of human response but me much better grounded than previous work.  It could really break down barriers to neuromodulation therapy.

Biological domain of the model
Lungs and heart
Structure(s) of interest in the model
Vagal and phrenic nerves
Spatial scales included in the model
0.0001 m to 0.01m
Time scales included in the model
0.001 s to 1 s
Other uses for the model (optional)

Utlimately, if there was some success to this approach, it may lead to a means of differentiating which animals are particularly relevant for certain neuromoduation therapies.

Additional comments about the model’s context (optional)

None

2. Data for building and validating the model
Data for building the model Published? Private? How is credibility checked? Current Conformance Level / Target Conformance Level
in vitro (primary cells cell, lines, etc.)
ex vivo (excised tissues)
in vivo pre-clinical (lower-level organism or small animal)
in vivo pre-clinical (large animal) Yes - feline, No - rat No -feline data is on Sparc Site, Yes - rat data is by collaborator Repeat experiments Adequate
Human subjects/clinical No Yes Population of patients Adequate
Other: ________________________ No, gross pathology of respective nerves Yes Multiple sample samples Adequate
Data for validating the model Published? Private? How is credibility checked? Current Conformance Level / Target Conformance Level
in vitro (primary cells cell, lines, etc.)
ex vivo (excised tissues)
in vivo pre-clinical (lower-level organism or small animal)
in vivo pre-clinical (large animal) see above see above Leave one out validation Adequate
Human subjects/clinical see above see above Leave one out validation Adequate
Other: ________________________
3. Validate within context(s)
Who does it? When does it happen? How is it done? Current Conformance Level / Target Conformance Level
Verification Users Matching of electrophysiological responses Basic correlation of electrophysiolological response adequate
Validation Users Novel electrophysiological responses in response to recorded stimulation Basic correlation of electrophysiolological response adequate
Uncertainty quantification Comes naturally out of analysis Every time Likely over a population of responses to similar disease adequate
Sensitivity analysis Users When the system is probed for parameter variations With variations in stimulation, what is the deviation from electrophysiological response adequate
Other:__________
Additional Comments
4. Limitations
Disclaimer statement (explain key limitations) Who needs to know about this disclaimer? How is this disclaimer shared with that audience? Current Conformance Level / Target Conformance Level
Until more data is acquired and analyzed with this framework in mind for development, it is likely that models built will be specific to this appliation. Research scientists primarily In publications Adequate
5. Version control
Current Conformance Level / Target Conformance Level
Partial
Naming Conventions? Repository? Code Review?
individual modeler Yes Yes No
within the lab NA NA NA
collaborators NA NA NA
6. Documentation
Current Conformance Level / Target Conformance Level
Code commented? Partial due to early development
Scope and intended use described? Partial due to early development
User’s guide? Partial due to early development
Developer’s guide? Insufficient due to early development
7. Dissemination
Current Conformance Level / Target Conformance Level
Adequate
Target Audience(s): “Inner circle” Scientific community Public
Simulations Yes
Models Yes
Software Yes
Results Yes
Implications of results Framework to create cross species clinical tools
8. Independent reviews
Current Conformance Level / Target Conformance Level
Insufficient, not built yet
Reviewer(s) name & affiliation: NA
When was review performed? NA
How was review performed and outcomes of the review? NA
9. Test competing implementations
Current Conformance Level / Target Conformance Level
Partial, not built yet but in addition to presented pipeline, we are exploring adaptive/adversarial machine learning methods that will provide parallel feedback mechanisms
Yes or No (briefly summarize)
Were competing implementations tested? NA - not built yet
Did this lead to model refinement or improvement? NA - not built yet
10. Conform to standards
Current Conformance Level / Target Conformance Level
Insufficient - not built yet
Yes or No (briefly summarize)
Are there operating procedures, guidelines, or standards for this type of multiscale modeling? No, but we do believe there will be
How do your modeling efforts conform? Given this is a very new build at an idea, we are not conforming to a large degree
11. (optional) Additional information to support items 1-10

This project is a new conceptualization for the use of preclinical systems.  To our knowledge, it has not been attempted.  If successful, the framework would be quite useful to the community.