Abstract
The study of neural population codes relies on massively parallel recordings in combination with theoretically motivated analysis tools. We applied two multi-site recording techniques to record from cells throughout cortical depth in a minimally invasive way. The feasibility of such experiments in area 17 of the anesthetized rat is demonstrated. Bayesian reconstruction and the interpretative framework of Fisher information are introduced. We demonstrate applicability and usefulness of Bayesian stimulus reconstruction and show that even small numbers of neurons can yield a high degree of representational accuracy under favorable conditions. Results are discussed and future lines of research outlined.
| Original language | English |
|---|---|
| Journal | Neurocomputing |
| Volume | 44-46 |
| Pages (from-to) | 407-416 |
| Number of pages | 10 |
| ISSN | 0925-2312 |
| DOIs | |
| Publication status | Published - 27.07.2002 |
Funding
Thanks are due to Sigrun Wicker and Barbara Klein for technical assistance. Multi-site recording probes were kindly provided by the University of Michigan Center for Neural Communication Technology sponsored by NIH/NCRR Grant P41 RR09754. Help and advice on the use of probes by Drs. Jamille Farraye Hetke, Chris Ellinger, Steven M. Bierer (University of Michigan) and Tim Blanche (University of British Columbia, Vancouver) have been invaluable for our experimental work. U. Hofmann was supported by a Feodor-Lynen-Fellowship and EU Grant IST-1999-10073, J.M. Bower, G.T.A. Kovacs, M.D. Hills and D.T. Kewley by NSF Grant #PC147129-1, W.A. Freiwald and H. Stemmann by a FNK Grant of the University of Bremen. Experimental and theoretical work was supported by SFB 517 “Neurocognition”.
