1 of 19

Pattern Separability Visual Feedback to Improve Pattern Recognition Decoding Performance

Gyorgy M. Levay, Richard Yang, Christopher Hunt, Megan Hodgson and Rahul Kaliki

​

Infinite Biomedical Technologies & Johns Hopkins University

​

​

October 1, 2022

2 of 19

Disclosures

Levay, Hodgson, and Kaliki are employees of Infinite Biomedical Technologies.

Pattern Recognition Visual Feedback

2

10/01/2022

3 of 19

Pattern Recognition Visual Feedback

3

10/01/2022

With traditional dual-site systems

CLOSE

OPEN

4 of 19

Pattern Recognition Visual Feedback

4

10/01/2022

With pattern recognition systems

5 of 19

Pattern Recognition Visual Feedback

5

10/01/2022

Ideal Calibration

Palm up

Hand close

Palm down

Hand open

6 of 19

Pattern Recognition Visual Feedback

6

10/01/2022

Realistic Calibration

Palm up

Hand close

Palm down

Hand open

7 of 19

Pattern Recognition Visual Feedback

7

10/01/2022

Improved Visibility – Current Solutions

8 of 19

Pattern Recognition Visual Feedback

8

10/01/2022

Improved Visibility – Proposed Solution

9 of 19

Pattern Recognition Visual Feedback

9

10/01/2022

Improved Visibility – Proposed Solution

10 of 19

Pattern Recognition Visual Feedback

10

10/01/2022

Study Purpose

To investigate the effects of the reviewer on various prosthesis control metrics during a Fitts Law Task.

11 of 19

Pattern Recognition Visual Feedback

11

10/01/2022

Experimental Protocol

Calibration

Control:

Virtual Arm Feedback

Experimental:

Reviewer Feedback

Fitts Law Task

Training Phase

12 of 19

Pattern Recognition Visual Feedback

12

10/01/2022

Experimental Setup

Day

Movements Calibrated

Total Classes

1-4

Rest, Power, Open, Pronation, Supination

5

5-7

Above + Tripod, Key

7

8-10

Above + Index Point, Pinch

9

13 of 19

Pattern Recognition Visual Feedback

13

10/01/2022

Fitts Law Task

14 of 19

Pattern Recognition Visual Feedback

14

10/01/2022

Results

15 of 19

Pattern Recognition Visual Feedback

15

10/01/2022

Results

16 of 19

Pattern Recognition Visual Feedback

16

10/01/2022

Results

Results

17 of 19

Conclusion

  • Reviewer showed improvement in all metrics
  • Training with reviewer has the potential to improve control

Pattern Recognition Visual Feedback

17

10/01/2022

18 of 19

Clinical Implications

  • Improve separability of movements
  • Explore options for additional distinct patterns
  • Transfer of knowledge for care teams
  • Reimbursement support

Pattern Recognition Visual Feedback

18

10/01/2022

19 of 19

Thank You!

This work was funded by the NIH National Institute of Neurological Disorders and Stroke

megan@i-biomed.com

Pattern Recognition Visual Feedback

19

10/01/2022