Abstract
We investigate population averaging as a preprocessing stage for linear FIR BMIs. Population averaging is a biologically-inspired technique based on spatial constraints and neuronal correlation. We achieve a statistically significant improvement in accuracy while substantially (45%) reducing model parameters. Further analysis is performed to show that population averaging improves model accuracy by reducing variance in estimating the firing rate from spike bins. However, we find that population averaging provides a greater accuracy improvement than other groupings which also reduce firing rate variance. Our results suggest that appropriate spatial organization of neural signals enhances BMI performance.
| Original language | English |
|---|---|
| Title of host publication | Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings |
| Pages | 1608-1611 |
| Number of pages | 4 |
| DOIs | |
| State | Published - Dec 1 2006 |
| Externally published | Yes |
| Event | 28th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS'06 - New York, NY, United States Duration: Aug 30 2006 → Sep 3 2006 |
Other
| Other | 28th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS'06 |
|---|---|
| Country | United States |
| City | New York, NY |
| Period | 8/30/06 → 9/3/06 |
Fingerprint
ASJC Scopus subject areas
- Bioengineering
Cite this
Improved linear BMI systems via population averaging. / DiGiovanna, Jack; Sanchez, Justin C.; Principe, Jose C.
Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2006. p. 1608-1611 4029847.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Improved linear BMI systems via population averaging
AU - DiGiovanna, Jack
AU - Sanchez, Justin C.
AU - Principe, Jose C.
PY - 2006/12/1
Y1 - 2006/12/1
N2 - We investigate population averaging as a preprocessing stage for linear FIR BMIs. Population averaging is a biologically-inspired technique based on spatial constraints and neuronal correlation. We achieve a statistically significant improvement in accuracy while substantially (45%) reducing model parameters. Further analysis is performed to show that population averaging improves model accuracy by reducing variance in estimating the firing rate from spike bins. However, we find that population averaging provides a greater accuracy improvement than other groupings which also reduce firing rate variance. Our results suggest that appropriate spatial organization of neural signals enhances BMI performance.
AB - We investigate population averaging as a preprocessing stage for linear FIR BMIs. Population averaging is a biologically-inspired technique based on spatial constraints and neuronal correlation. We achieve a statistically significant improvement in accuracy while substantially (45%) reducing model parameters. Further analysis is performed to show that population averaging improves model accuracy by reducing variance in estimating the firing rate from spike bins. However, we find that population averaging provides a greater accuracy improvement than other groupings which also reduce firing rate variance. Our results suggest that appropriate spatial organization of neural signals enhances BMI performance.
UR - http://www.scopus.com/inward/record.url?scp=34047177010&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=34047177010&partnerID=8YFLogxK
U2 - 10.1109/IEMBS.2006.260496
DO - 10.1109/IEMBS.2006.260496
M3 - Conference contribution
C2 - 17946469
AN - SCOPUS:34047177010
SN - 1424400325
SN - 9781424400324
SP - 1608
EP - 1611
BT - Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings
ER -