de.mpg.escidoc.pubman.appbase.FacesBean
English
 
Help Guide Disclaimer Contact us Login
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Poster

Splitting local ensembles by spike-sorting degrades the quality of neuronal object representations in macaque prefrontal cortex

MPS-Authors
http://pubman.mpdl.mpg.de/cone/persons/resource/persons84293

Waizel,  M
Department Physiology of Cognitive Processes, Max Planck Institute for Biological Cybernetics, Max Planck Society;

http://pubman.mpdl.mpg.de/cone/persons/resource/persons84098

Munk,  MJH
Department Physiology of Cognitive Processes, Max Planck Institute for Biological Cybernetics, Max Planck Society;

Locator
There are no locators available
Fulltext (public)
There are no public fulltexts available
Supplementary Material (public)
There is no public supplementary material available
Citation

Waizel, M., Staedtler ES, Pipa G, Chen, N., & Munk, M. (2007). Splitting local ensembles by spike-sorting degrades the quality of neuronal object representations in macaque prefrontal cortex. Poster presented at 37th Annual Meeting of the Society for Neuroscience (Neuroscience 2007), San Diego, CA, USA.


Cite as: http://hdl.handle.net/11858/00-001M-0000-0013-CB49-9
Abstract
As most cortical neurons are broadly tuned to various stimulus parameters, accurate representations of individual objects in short-term memory that can support the monkey’s decision need to be based on the activity of neuronal ensembles. We have in the past (Städtler et al., SfN 2006) analyzed simultaneously recorded multi-unit signals and found to our surprise that stimulus selectivity at individual sites was higher than what had been published for single units. We had expected that a signal that consists of the activity of several neurons should provide less selectivity because combining broadly tuned activity should result in even broader tuning by blurring. In order to test whether splitting up the multi-unit signal improves or degrades stimulus selectivity, we sorted multi-unit activity from 97 sites by means of semi-automatic PCA-based clustering (Chen et al., in prep.) into 413 sorted units. As for the multi-unit analysis, we calculated one-way ANOVAs (p<0.01) to detect stimulus selectivity and determined stimulus specificity by subsequent posthoc comparison (Scheffé). Only 11.6 of responsive sorted-units exhibited object-selective responses compared to 18.6 of responsive multi-units. Even more detrimental was the effect of sorting on stimulus specificity: only 2.9 of sorted units allowed for the discrimination of objects compared to 7.2 of the unsorted multi-unit signals. If this result is not due to a problem with detectability of weak signals, it suggests that the selectivity of local ensembles might be caused by coordination of their members. However, before we can directly address this possibility we need to improve our spike sorting methods such that temporally overlapping spike waveforms can unequivocally assigned to subsets of local neurons. For the time being we asked whether selectivity and specificity of distributed activity patterns based on sorted units combined across sites would profit from the contribution of a larger number of units, the prediction being that more independent signals could in principle provide more differentiating patterns. We therefore calculated binary activity patterns derived from epochs with a significant difference in firing rate compared to baseline. Distributed patterns of sorted activity provided a lower percentage of distinguishable patterns than the same analysis revealed for multi-unit signals. These results show that decomposition of local ensemble activity has direct impact on the quality of stimulus representations which in turn suggests that stimulus coding in the short-term memory of prefrontal cortex seems to depend on the coordinated activity of neuronal populations.