DOC PREVIEW
Stimulus-invariant processing and spectrotemporal reverse correlation in primary auditory cortex

This preview shows page 1-2-3-24-25-26 out of 26 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 26 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 26 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 26 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 26 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 26 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 26 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 26 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

J Comput Neurosci (2006) 20:111–136DOI 10.1007/s10827-005-3589-4Stimulus-invariant processing and spectrotemporal reversecorrelation in primary auditory cortexDavid J. Klein · Jonathan Z. Simon ·Didier A. Depireux · Shihab A. ShammaReceived: 29 September 2004 / Revised: 5 June 2005 / Accepted: 12 July 2005 / Published online: 20 February 2006CSpringer Science + Business Media, LLC 2006Abstract The spectrotemporal receptive field (STRF) pro-vides a versatile and integrated, spectral and temporal, func-tional characterization of single cells in primary auditorycortex (AI). In this paper, we explore the origin of, and rela-tionship between, different ways of measuring and analyz-ing an STRF. We demonstrate that STRFs measured using aspectrotemporally diverse array of broadband stimuli—suchas dynamic ripples, spectrotemporally white noise, and tem-porally orthogonal ripple combinations (TORCs)—are verysimilar, confirming earlier findings that the STRF is a robustlinear descriptor of the cell. We also present a new deter-ministic analysis framework that employs the Fourier se-ries to describe the spectrotemporal modulations containedin the stimuli and responses. Additional insights into theAction Editor: Mathew WienerD. J. Klein () · D. A. Depireux · S. A. ShammaInstitute for Systems Research, University of Maryland,College Park, MD 20742, USAe-mail: [email protected]. J. Klein · J. Z. Simon · S. A. ShammaDepartment of Electrical and Computer Engineering, Universityof Maryland,College Park, MD 20742, USAJ. Z. SimonDepartment of Biology, University of Maryland,College Park, MD 20742, USAD. A. DepireuxDepartment of Anatomy and Neurobiology, University ofMaryland,Baltimore, MD 21201, USAD. J. KleinInstitute for Neuroinformatics, University/ETH Z¨urich,8057 Z¨urich, SwitzerlandSTRF measurements, including the nature and interpreta-tion of measurement errors, is presented using the Fouriertransform, coupled to singular-value decomposition (SVD),and variability analyses including bootstrap. The results pro-mote the utility of the STRF as a core functional descriptorof neurons in AI.Keywords spectrotemporal receptive field.modulationtransfer function.auditory cortex.ripple.variability.singular-value decomposition.ferret1. IntroductionIt has been over twenty years since the spectrotemporal re-ceptive field (STRF) was conceived to describe and measureauditory neurons’ joint sensitivity to the spectral and tem-poral dimensions of acoustical energy (Hermes et al., 1981;Aertsen and Johannesma, 1981b; Smolders et al., 1979;Eg-germont et al., 1981; Johannesma and Eggermont, 1983).It was specifically associated with (1) stimuli characterizedby randomly varying spectrotemporal features, and (2) anapproach labeled reverse correlation, by which the neuroninforms the experimenter, via action potentials, of the fea-tures that were of interest to it (de Boer and de Jongh, 1978;Eggermont et al., 1983b). The STRF offered a view of neu-ronal function that complemented, and was usually consis-tent with, that obtained using classical stimuli such as tones(tuning curves and rate-level functions), clicks (impulse re-sponses), and noise (bandwidth sensitivity). In addition, itneatly fit within an analytical framework, bolstered by thefields of time-frequency analysis (Cohen, 1995) and non-linear systems theory (Eggermont, 1993), within which thefunctionality of neurons could, in principle, be systemati-cally explored to any level of detail.Springer112 J Comput Neurosci (2006) 20:111–136The term “STRF” does not denote here the full com-plex (likely nonlinear) receptive field of an auditory neu-ron. Rather it is a technical term that has traditionallybeen used to refer specifically to the linear relationship be-tween the time-dependent spike rate of a neuron and thetime- and frequency-dependent energy—in short, the dy-namic spectrum—of a stimulus. In order to measure theSTRF, the reverse-correlation approach prescribes comput-ing the average dynamic spectrum of those portions of astimulus preceding the neuron’s spikes. In this context, theSTRF is commonly interpreted as the spectrotemporal pat-tern that optimally activates a neuron (Young, 1998). The-oretically, as long as all patterns occur randomly, indepen-dently, and equiprobably, the STRF can be revealed by this“spike-triggered average” (Eggermont, 1993).As with tuning curves,rate-level functions, and other com-monly used neuronal response measures, the STRF providesonly a limited view of the receptive field of a neuron, onethat is useful only within the context of the experiment orthe nature of information sought from it. For example, tun-ing curves are useful as approximate indicators of a unitsBF and bandwidth, but are largely irrelevant as a gauge ofits dynamic range and temporal properties. Similarly, theSTRF is a useful measure of spectrotemporal features likelyto drive a cells responses. However, being a measure of thelinear component of the stimulus-response relationship, it ismostly effective in predicting the linear aspects of the re-sponses, predictions that can be accurate if the non-linearportions are small or are well known and can be accountedfor in the measurement (e.g., spike-rate rectification and satu-ration). In some cases, the linear component of the responseis small and hence one does not expect clean and reliableSTRF measurements, i.e., the STRFs exhibit significant ran-domness or high variability across presentations, or are poorpredictors of responses to novel stimuli. Examining thesesources of variability and prediction errors provides usefulinformation regarding the limitations of the STRF and waysto extend it beyond the linear domain.Although the STRF has been slow to mature, it is nowincreasingly used to study the physiology of central auditoryneurons. In retrospect, the often slow pace of progress can bepartially attributed to the reverse-correlation methodology,which remains fairly opaque. In particular, reverse correla-tion provides no straightforward formal basis for describingthe effectiveness of, or relations between, specific stimuli, be-cause only the average statistics of stimuli are specified. Forexample, Gaussian broad-band noise, the “ideal” stimulus forreverse-correlation, is often ineffective when applied to cen-tral auditory neurons (but see (Keller and Takahashi, 2000)).Meanwhile, a range of other stimuli and associated tech-niques have been auditioned, modulated


Stimulus-invariant processing and spectrotemporal reverse correlation in primary auditory cortex

Download Stimulus-invariant processing and spectrotemporal reverse correlation in primary auditory cortex
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Stimulus-invariant processing and spectrotemporal reverse correlation in primary auditory cortex and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Stimulus-invariant processing and spectrotemporal reverse correlation in primary auditory cortex 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?