Sensory adaptation


adaptation occurs in a variety of forms in all centripetal systems, motivating the question : what is its function ? A generative approach path has been to hypothesize that adaptation helps neural systems to efficiently encode stimuli whose statistics vary in meter. To encode efficiently, a neural system must change its code strategy, or calculation, as the distribution of stimulation change. Information theoretical methods allow this effective coding hypothesis to be tested quantitatively. empirically, adaptive processes occur over a across-the-board roll of timescales. On short timescales, underlying mechanisms include the contribution of intrinsic nonlinearities. Over longer timescales, adaptation is often power-law-like, implying the coexistence of multiple timescales in a single adaptive work. Models demonstrate that this can result from mechanisms within a single nerve cell .

Adaptation as efficient coding

Barlow ’ s effective coding guess suggests that, given a finite capability to transmit information, neural systems employ an optimally effective coding scheme to represent the inputs that they typically process [ 1 – 3, Box 1 ]. however, when collected over relatively short time or distance scales, the local statistics of many natural stimuli differ greatly from their global distribution. For exemplar, luminosity and contrast in natural ocular environments vary over orders of magnitude across prison term in a day or across space in a building complex scene. A sensory system that matched its distribution of outputs to the ball-shaped distribution of stimuli would then be inefficient in transmitting the stimulation ’ local distribution. Under these circumstances, one might expect the organization ’ sulfur coding strategy to adapt to local anesthetic characteristics of the stimulation statistics .

Efficient coding (Box 1)

Maximizing efficiency with a fasten dynamic rate requires that the organization maps its inputs to its outputs such that all outputs are evenly likely [, 25, 64 ]. The optimum code scheme thus depends on the statistics of the stimulation that the system represents.

An external file that holds a picture, illustration, etc.
Object name is nihms32849f1.jpgOpen in a separate window recent experimental designs have begun to test the effective coding guess in the context of adaptation. In these experiments, the distribution of a random time-varying stimulation, rather than a single stimulation parameter, is varied. This design allows one to examine the tease strategy that the neural system uses to represent an entire stimulation distribution, and to relate changes in coding strategy to changes in stimulation distribution. The model underlying these experiments presents the undertaking of adaptation as basically an inference trouble : the timescale for the switch in coding scheme can not be shorter than the time required for the system to “ teach ” the new distribution. In club to analyze experiments of this type, one must describe the coding scheme of the system during changes in the stimulation distribution. To do so, it is necessity to reduce the system ’ randomness entire input-output mapping to a simple portrayal. Linear-nonlinear ( LN ) models have much been successful in capturing changes in the calculation of an adapt system ( Box 2 ) .

Capturing adaptive computation (Box 2)

To efficaciously characterize adaptation, one must begin with a portrayal of calculation. A knock-down method acting for characterizing nervous calculation is to approximate the neural system as inaugural linearly filtering the stimulation by an identified relevant feature or set of features, and then to generate spikes according to a nonlinear officiate of the stimulation ’ similarity to the feature ( randomness ) ( ). Models of this simple type are known as linear/nonlinear ( LN ) models, and have had considerable success in capturing some aspects of neural processing [ 26 ]. In this simplify framework, adaptation to stimulus statistics might affect the features that linearly filter the stimulation, or the nonlinear function that determines the probability to fire ( ). A change in reply gain ( without a deepen in the feature ) can be manifested as either a upright scale of the linear sport or a horizontal scale of the nonlinearity. Changes in the feature of speech for unlike stimulation conditions occur, for case, in the retina ; in moo inner light levels, the centripetal fields of retinal ganglion cells show increase temporal integration and decreased inhibitory surround [ 25 ]. One can sample these potentially time-dependent components using reverse correlation either with spikes conditioned on the phase of their arrival prison term with esteem to the stimulation hertz if the stimulation is changing sporadically, or with an adaptive filter [ 65 ] .

Systems adapt to a variety of stimulus statistics

The simplest instantiation of a stimulation probability distribution is one in which the stimulation takes one of two possible values. If one of these values is presented more frequently than the early, a system may adapt to give a stronger answer to the rare stimulation. This effect was observed in computerized tomography auditory lens cortex A1 in response to two tones presented with different probabilities [ 4, 5 ]. A potential substrate for this effect was demonstrated in a culture network by Eytan et aluminum. [ 6 ], who varied the relative frequency of stream injection inputs applied at different locations. such stimulus-specific adaptation may implement a kind of knickknack signal detection, in which the persuasiveness of the response is adjusted according to the information it carries. More broadly, the effective coding guess might be taken to suggest that stimulation encoding is sensitive to the variations in stimulation statistics seen in natural stimuli. The properties of natural ocular scenes, in detail, have been extensively studied [ 7 ]. A dim-witted model for natural stimuli is of local gaussian fluctuations with a long-tailed mean and discrepancy modulated on longer time or distance scales [ 8 ]. Inspired by this description of natural stimuli, a class of experiments has examined adaptation to a change in the mean or variance of a gaussian white noise stimulation. Using a switching paradigm in which a random stimulation is chosen from a distribution whose parameters change sporadically between two values ( ), Smirnakis et alabama. [ 9 ] showed that retinal ganglion cells exhibit an adaptive change in firing pace when the variability of a flicker light stimulation changes. In the fly movement sensitive nerve cell H1, Brenner et alabama. [ 10 ] computed an LN model for unlike variances of a randomly vary speed stimulation. They showed that the nonlinear amplification function adapted such that the scaling of the stimulation axis was normalized by the stimulation standard deviation, and that this serves to maximize information infection about the stimulation. further, this addition change occurs in ~100 mississippi, quickly maximizing data transmittance during continuous changes in stimulation variation [ 11 ]. A classify slower adaptive process modulates the overall fire pace on much longer timescales. analogously, retinal ganglion cells display contrast gain master [ 12 – 16 ], which occurs on a much faster timescale than rate changes due to contrast adaptation [ 9, 14 ] .An external file that holds a picture, illustration, etc.
Object name is nihms32849f2.jpgOpen in a separate window adaptation to stimulus variation has besides been observed in several higher mind regions. In scab barrel cerebral cortex, when the variance of a white randomness motion of the whiskers was changed, the relevant features remained approximately unaltered, but the gain curves showed a change in scale by the stimulation standard diversion [ 17 ]. In discipline L, the avian analogue of primary auditory cortex, Nagel and Doupe [ 18 ] observed rapid changes in filters and advance curves a well as a slower transition of the overall fire pace, similar to observations in H1 and RGCs, as the distribution of sound intensity was varied. While individual neurons in inferior colliculus responding to sound amplitudes showed a diverseness of response changes during adaptation, Dean et alabama. [ 19 ] used Fisher data to demonstrate that the population as a hale shifted responses to best encode the gamey probability sounds, even when the distributions were relatively complex, such as bimodal. nervous responses in macaque subscript worldly cortex adapted to the width of the distribution of prototype stimuli along randomly chosen stimulation directions [ 20 ] .

Beyond white noise

In the LGN, Mante et aluminum. [ 21 ] examined the interaction between adaptation to the base and the discrepancy ( mean luminosity and contrast ) of drifting grating stimuli and found that the changes in filters ascribable to luminance and contrast adaptation are freelancer. suggestively, their analysis of natural images besides showed independence of luminosity and line. The lapp group found no adaptation to higher ordain statistics—the lopsidedness and kurtosis—of a random checkboard stimulation in LGN [ 22 ]. however, Hosoya et aluminum. [ 23 ] generalized a former find of rate adaptation to the spatial scale of a flickering checkerboard [ 9 ] to demonstrate adaptation to a variety of arbitrary spatiotemporal correlations in ocular stimulation in RGCs, and showed that the new filters that evolve after exposure to these correlations act to remove the correlations and then perform predictive coding [ 24, 25 ]. Determining the effect of complex changes in stimulation distribution is unmanageable to address due to the biases introduced into flannel randomness analysis by non-Gaussian stimuli [ 26 ], which are unmanageable to separate from ascertained dependences of sample centripetal fields on the stimulation corps de ballet [ 27 – 30 ]. Sharpee et alabama. [ 31 ] introduced an information theoretical overrule correlation coefficient method which finds the stimulation dimensions that maximize reciprocal information between spiking responses and the stimulation. This method was used to find significant differences between the features encoded by V1 neurons in a white noise ensemble and a natural stimulation ensemble [ 32 ] .

Multiple timescales

How might the goal of maintaining efficient information transmission constrain the dynamics of adaptation ? In tracking changing stimulation statistics, there are two relevant timescales for any system : the feature timescale of changes in the stimulation distribution, and the minimum meter required by an ideal perceiver to estimate the parameters of the new distribution. The first timescale is established by the environment, while the second is determined by statistics and sets a lower bind on how promptly any adjust organization could estimate parameters of the new distribution. Given these constraining timescales, an adapting system should choose an allow estimate timescale for computing local stimulation statistics. For exercise, consider a system that adapts to the local stimulation mean. If this system estimates the local mean by averaging over only a few samples, the system would amplify noise and transmit little information about its stimulation [ 33 ]. conversely, a system that averages over a timescale much longer than the timescale of changes in stimulation mean will not be optimally adapted to the local stimulation ensemble. This argument assumes that the nervous organization can choose an adaptation timescale to match the dynamics of stimulation statistics. experimentally, it appears that two dissociable phenomena describe the dynamics of adaptation to variation or contrast, at least in early ocular and auditory systems. As discussed above, the first of these components quickly rescales the system ’ randomness input-output gain following a change in stimulation statistics [ 11 – 16, 18, 34 ]. The moment component of adaptation dynamics is a behind change in the system ’ s base firing rate [ 9, 11, 13, 15, 18, 35, 36 ]. It is not even clear what, if any, relationship exists between these two phenomena. In the fly H1 nerve cell, these effects appear to be freelancer [ 11 ], but such a consequence does not appear necessary a priori. At least in some cases, the adaptive rescale of input-output functions in rat barrel cerebral cortex appears to follow the dull timescale rate adaptation [ 17 ]. possibly coherent with this, Webber and Stanley found that transient and steady-state adaptation in this area could be modeled with a unmarried express varying [ 37 ]. The timescale of the flying gain exchange is often on the rate of the timescale of the arrangement ’ s relevant feature, and does not appear to depend on the timescale of switches in stimulation ensemble. In at least some cases, fast amplification changes in answer to changes in stimulation variability or correlation coefficient time are a consequence of the arrangement ’ s inactive nonlinearity with no change in system parameters [ 38 – 44 ]. Whether such effects should rightly be called adaptation is something of a philosophic question. There is no doubt that finite dimensional stimulus/response characterizations such as LN models are limited and adaptation may appear to change exemplar parameters. On the other hand, these models may, in some cases, capture all stimulation dependence if appropriately extended. Slower changes in overall excitability have variable dynamics and may be subserved by a wide assortment of mechanisms. Spike frequency adaptation ( SFA ) occurs over many timescales in cortical neurons [ 45 ], and has been analytically described for childlike model neurons [ 46, 47 ]. By specifying only the initial and steady-state fire rate-input ( f-I ) crook angstrom well as the effective time constant, SFA can be described without cognition of especial neural dynamics [ 48, 49 ]. dull currents have been implicated in altering the gain of f-I curves, allowing neurons to remain sensitive to input fluctuations at high beggarly currents [ 50 – 52 ] .

Power law adaptation

In some cases, the dynamics of slow changes in excitability might be matched to the dynamics of stimulation changes. many researchers using a switching paradigm with a single switching timescale have reported the dynamics of this decelerate addition change as an exponential march with a fixed time scale [ 5, 9, 13 – 15, 35 ]. however, in the fly ocular nerve cell H1, when the time between stimulation changes was varied, the apparent adaptation timescale scaled proportionately [, 11, 53 ]. therefore, the dynamics of slow advance changes in tent-fly H1 are consistent with a power-law preferably than an exponential march [ 36, 54 ]. Power police dynamics are significant because there is no inside timescale : dynamics are invariant with respect to changes in temporal scale, and such a system could therefore adjust its effective adaptation timescale to the environment .An external file that holds a picture, illustration, etc.
Object name is nihms32849f3.jpgOpen in a separate window Although studies in other systems have not explicitly tested for multiple clock time scales, results from studies of temporal contrast adaptation in poker and mammal ( rabbit and guinea hog ) retina suggest that the apparent time constant of the boring gain change indeed varies as a serve of the period between stimulation switches ( ). few studies provide direct attest for the biophysical mechanisms underlying multiple timescale dynamics. Power police dynamics can be approximated by a cascade of many exponential processes [ 36, 54, 55 ]. therefore, a leading hypothesis is that multiple timescale dynamics are the solution of a cascade of exponential processes in a cell or network. multiple timescales exist in the numerousness of channel dynamics present in one neurons [ 45 ]. even in single channels, power-law recovery from inactivation has been shown in disjunct NaII and NaIIa channels [ 56 ]. This demeanor is captured by a stochastic sodium channel model that includes a Markov chain of multiple inactivation states [ 55 ] .

Intrinsic properties or circuit mechanisms?

A leading campaigner for a mechanism of contrast amplification control in V1 is dissentious standardization, in which the output of a given nerve cell is modulated by feedback from the responses of neurons with similar centripetal fields [ 57, 58 ]. however, many of the mechanism we have discussed here may operate at the level of one neurons [ 59 ]. holocene solve has made considerable build up in elucidating where in particular circuits adaptation occurs. In salamander retinene ganglion cells, rapid contrast adaptation is partially inherited from the adaptation of synaptic inputs [ 13, 60 ] while a moment part is contributed by intrinsic mechanisms [ 61 ]. Manookin et aluminum. [ 35 ] besides find that recovery from high contrast stimulation in guinea pig RGCs, characterized by a slow “ afterhyperpolarization, ” is mediated in large depart by inherit changes in synaptic inputs with an extra intrinsic component. In mouse retina, adaptation to dim mean background luminosity occurs in rod photoreceptors and at the perch bipolar-to-AII amacrine cell synapse [ 62 ]. In this encase, the dominant web site of adaptation was predicted by the probably site of impregnation in reception due to overlap of signals in the retinene circuitry. In rat barrel cerebral cortex, Katz et alabama. [ 63 ] showed that a subthreshold component of adaptation is whisker-specific, while responses in barrel cortex are multi-whisker, implying that the adaptation occurs in intracortical or thalamocortical connections as opposed to intrinsic mechanisms in the barrel cortical neurons .


A growing body of evidence suggests that representations at all levels throughout centripetal serve pathways are credit card, depending on the recent history of the stimulation, on a range of timescales varying from about instantaneous to timescales more typically associated with synaptic changes. This malleability can increase the information infection rate of the signal. One would frankincense like to determine whether constraints on timescales are imposed by the time required to learn dynamic effective representations. It is becoming clear up that some components of what we think of as advance march may be occurring at low levels. Furthermore, some types of sophisticated apparent learning effects may be a result of intrinsic nonlinearities. A see of sensory systems as a simple feed-forward relay of percolate centripetal data from transducers to cortex is no longer allow. rather, we must consider the statistics of the natural worldly concern, malleability at multiple levels of sensational process, and the consequences for encoding of sensory information at each stage .


We would like to thank Felice Dunn, Gabe Murphy, Fred Rieke and Rebecca Mease for utilitarian comments and discussions. This employment was supported by a Burroughs-Welcome Careers at the Scientific Interface grant ( AF, BW and BL ), NIH T32EY-0731 ( BW ), NIH MSTP T32-07266 ( BL ), ARCS ( BL ) .


Annotations Sharpee et al., 2006 :

This newspaper applies a novel information theoretical invert correlation method acting [ 31 ] to obtain kat V1 centripetal fields during screening of a white noise stimulation and of natural movies. The centripetal fields differed in the two cases in their low-pass properties, in such a way as to transmit like low frequency power in the two cases. The information transmitted about the stimulation was determined to change on the order of ~100 sec. Maravall et al., 2006 : It is shown that adaptation to stimulus variance besides occurs in neurons in primary somatosensory cerebral cortex ; here, rat barrel lens cortex. Due to local anesthetic phase invariability of the reply, features and profit curves were extracted using covariance analysis. While features did not change with variance, the ranges of the acquire curves were found to scale proportionately to the stimulation standard deviation in the majority of neurons, maintaining data transmission. Nagel and Doupe, 2006 : The responses of neurons in avian field L are analyzed during changes in mean and variance of the randomly varying amplitude of a broadband noise input signal. The filters of an LN model change systematically with changes in the mean. As in [ 11 ] and [ 14 ], while the firing pace changes slowly following a exchange in stimulation discrepancy, the filters and derive curves change about instantaneously. Gilboa, Chen, and Brenner, J Neurosci 2005 : Following from the observation of Toib, Lyakhov, and Marom [ 56 ] that mammalian brain sodium channels display power-law like recovery from inactivation, this article models sodium channel deactivation using a Markov chain of one energizing state and many deactivation states, where neural excitability is related to the fraction of available channels. The feature time scale of adaptation is found to depend on stimulation duration by a power-law scale. The key expression of this model is that the clock scale of recovery depends on how its large pool of devolve inactive states is populated, which in turn is affected by stimulation history. Borst, Flanagin, and Sompolinsky, PNAS, 2005 : In the fly ocular system, previous work has demonstrated that the H1 nerve cell adapts its input/output relationship as stimulation statistics change. These authors find that this apparently complicated adaptation can be explained by the intrinsic properties of a Reichhardt correlator motion detector, where such rescale is a consequence of the system nonlinearity acting on a multidimensional stimulation. This acquire change appears in a reduce dimensional exemplar as a result of early, correlated, dimensions of the stimulation that are not included in the LN model yet affect the response differently at different variances. Hosoya et aluminum. ( 2006 ) : This study explores adaptation to complex spatial or temporal correlations in poker and rabbit retina, showing that the receptive fields of many RGCs change following foreplay with these correlated stimuli such that the response to novel stimulation is enhanced compared to the reception to the adapting stimulation. The authors reject the hypothesis of fatigue duty of convention matching cells in favor of a model of Hebbian malleability at inhibitory amacrine-to-RGC synapses. This stopping point is intriguingly at odds with other studies that find adaptation in the retina to simpler correlations ( mean, discrepancy ) in the absence of inhibitory transmission [ 34 ]. Manookin and Demb ( 2006 ) :

This study adds to our understanding of the relationship between network and intrinsic components of temporal contrast adaptation in retina. The authors conclude that recovery from a senior high school contrast stimulation, characterized by a behind afterhyperpolarization in guinea pig RGC membrane potential and firing rate, is inherited largely from presynaptic mechanisms. In addition, they find a small contribution of intrinsic mechanisms in the RGC. This work is consistent with previous findings in salamander retina [ 13, 61 ]. Publisher’s Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early translation of the manuscript. The manuscript will undergo copyread, typeset, and follow-up of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal refer .

reservoir :
Category : Tech

About admin

I am the owner of the website, my purpose is to bring all the most useful information to users.

Check Also


Manage participants in a zoom meeting webinar

Call the people who attend the meet as follows Alternate host host Who scheduled the …

Leave a Reply

Your email address will not be published.