Ultra-fine frequency tuning revealed in single neurons of human auditory cortex.

Just-noticeable differences of physical parameters are often limited by the resolution of the peripheral sensory apparatus. Thus, two-point discrimination in vision is limited by the size of individual photoreceptors. Frequency selectivity is a basic property of neurons in the mammalian auditory pathway. However, just-noticeable differences of frequency are substantially smaller than the bandwidth of the peripheral sensors. Here we report that frequency tuning in single neurons recorded from human auditory cortex in response to random-chord stimuli is far narrower than that typically described in any other mammalian species (besides bats), and substantially exceeds that attributed to the human auditory periphery. Interestingly, simple spectral filter models failed to predict the neuronal responses to natural stimuli, including speech and music. Thus, natural sounds engage additional processing mechanisms beyond the exquisite frequency tuning probed by the random-chord stimuli.

Sound-localization experiments with barn owls in virtual space: influence of interaural time difference on head-turning behavior.

Specific cues in a sound signal are naturally linked to certain parameters in acoustic space. In the barn owl, interaural time difference (ITD) varies mainly with azimuth, while interaural level difference (ILD) varies mainly with elevation. Previous data suggested that ITD is indeed the main cue for azimuthal sound localization in this species, while ILD is an important cue for elevational sound localization. The exact contributions of these parameters could be tested only indirectly because it was not possible to generate a stimulus that contained all relevant spatial information on the one hand, and allowed for a clean separation of these parameters on the other hand. Virtual auditory worlds offer this opportunity. Here we show that barn owls responded to azimuthal variations in virtual space in the same way as to variations in free-field stimuli. We interpret the increase of turning angle with sound-source azimuths (up to +/- 140 degrees) such that the owls did not experience front/back confusions with virtual stimuli. We then separated the influence of ITD from the influence of all other stimulus parameters by fixing the overall ITD in virtual stimuli to a constant value (+100 micros or +100 micros) while leaving all other sound characteristics unchanged. This manipulation influenced both azimuthal and elevational components of head arms. Since the owls' azimuthal head-turn amplitude always resembled the value signified by the ITD, these data demonstrated that azimuthal sound localization is influenced only by ITD both in the frontal hemisphere and in large parts of the rear hemisphere. ILDs did not have an influence on azimuthal components of head turns. While response latency to normal virtual stimuli was found to be largely independent of stimulus position, response delays of the head turns became longer if the ITD information pointed into a different hemisphere as the other cues of the sounds.

Recurrence methods in the analysis of learning processes.

The goal of most learning processes is to bring a machine into a set of "correct" states. In practice, however, it may be difficult to show that the process enters this target set. We present a condition that ensures that the process visits the target set infinitely often almost surely. This condition is easy to verify and is true for many well-known learning rules. To demonstrate the utility of this method, we apply it to four types of learning processes: the perceptron, learning rules governed by continuous energy functions, the Kohonen rule, and the committee machine.

Auditory edge detection: a neural model for physiological and psychoacoustical responses to amplitude transients.

Primary segmentation of visual scenes is based on spatiotemporal edges that are presumably detected by neurons throughout the visual system. In contrast, the way in which the auditory system decomposes complex auditory scenes is substantially less clear. There is diverse physiological and psychophysical evidence for the sensitivity of the auditory system to amplitude transients, which can be considered as a partial analogue to visual spatiotemporal edges. However, there is currently no theoretical framework in which these phenomena can be associated or related to the perceptual task of auditory source segregation. We propose a neural model for an auditory temporal edge detector, whose underlying principles are similar to classical visual edge detector models. Our main result is that this model reproduces published physiological responses to amplitude transients collected at multiple levels of the auditory pathways using a variety of experimental procedures. Moreover, the model successfully predicts physiological responses to a new set of amplitude transients, collected in cat primary auditory cortex and medial geniculate body. Additionally, the model reproduces several published psychoacoustical responses to amplitude transients as well as the psychoacoustical data for amplitude edge detection reported here for the first time. These results support the hypothesis that the response of auditory neurons to amplitude transients is the correlate of psychoacoustical edge detection.

Synthesizing spatially complex sound in virtual space: an accurate offline algorithm.

The study of spatial processing in the auditory system usually requires complex experimental setups, using arrays of speakers or speakers mounted on moving arms. These devices, while allowing precision in the presentation of the spatial attributes of sound, are complex, expensive and limited. Alternative approaches rely on virtual space sound delivery. In this paper, we describe a virtual space algorithm that enables accurate reconstruction of eardrum waveforms for arbitrary sound sources moving along arbitrary trajectories in space. A physical validation of the synthesis algorithm is performed by comparing waveforms recorded during real motion with waveforms synthesized by the algorithm. As a demonstration of possible applications of the algorithm, virtual motion stimuli are used to reproduce psychophysical results in humans and for studying responses of barn owls to auditory motion stimuli.

Relating cluster and population responses to natural sounds and tonal stimuli in cat primary auditory cortex.

Most information about neuronal properties in primary auditory cortex (AI) has been gathered using simple artificial sounds such as pure tones and broad-band noise. These sounds are very different from the natural sounds that are processed by the auditory system in real world situations. In an attempt to bridge this gap, simple tonal stimuli and a standard set of six natural sounds were used to create models relating the responses of neuronal clusters in AI of barbiturate-anesthetized cats to the two classes of stimuli. A significant correlation was often found between the response to the separate frequency components of the natural sounds and the response to the natural sound itself. At the population level, this correlation resulted in a rate profile that represented robustly the spectral profiles of the natural sounds. There was however a significant scatter in the responses to the natural sound around the predictions based on the responses to tonal stimuli. Going the other way, in order to understand better the non-linearities in the responses to natural sounds, responses of neuronal clusters were characterized using second order Volterra kernel analysis of their responses to natural sounds. This characterization predicted reasonably well the amplitude of the response to other natural sounds, but could not reproduce the responses to tonal stimuli. Thus, second order non-linear characterizations, at least those using the Volterra kernel model, do not interpolate well between responses to tones and to natural sounds in auditory cortex.

Responses to linear and logarithmic frequency-modulated sweeps in ferret primary auditory cortex.

Multi-unit responses to frequency-modulated (FM) sweeps were studied in the primary auditory cortex of ferrets using six different stimulation paradigms. In particular, the differences between the responses to linear FM sweeps (where frequency changes linearly with time) and logarithmic FM sweeps (where frequency changes exponentially with time) were emphasized. Some general features of the responses to FM sweeps are independent of the exact details of the frequency trajectory. Both for linear and for logarithmic FM sweeps, a short burst of spikes occurred when the sweep reached a triggering frequency close to the best frequency of the cluster. The neuronal preference for FM velocity was also independent of frequency trajectory. Thus, clusters that responded best to slow logarithmic FM also preferred slow linear FM and vice versa. Consequently, topographic distributions of velocity preference were roughly independent of the stimulation paradigm. Other characteristics of the responses, however, depended on the exact details of the frequency trajectory. A significant number of clusters showed large differences in directional sensitivity between linear and logarithmic FM sweeps; these differences depended on the velocity preference of the clusters in some paradigms but not in others. Consequently, topographic distributions of directional sensitivity differed between linear and logarithmic paradigms. In conclusion, some characteristics of cluster responses to FM sweeps depend on the exact details of the stimulation paradigm and are not 'invariants' of the cluster.

Spectral integration by type II interneurons in dorsal cochlear nucleus.

The type II unit is a prominent inhibitory interneuron in the dorsal cochlear nucleus (DCN), most likely recorded from vertical cells. Type II units are characterized by low rates of spontaneous activity, weak responses to broadband noise, and vigorous, narrowly tuned responses to tones. The weak responses of type II units to broadband stimuli are unusual for neurons in the lower auditory system and suggest that these units receive strong inhibitory inputs, most likely from onset-C neurons of the ventral cochlear nucleus. The question of the definition of type II units is considered here; the characteristics listed in the preceding text define a homogeneous type II group, but the boundary between this group and other low spontaneous rate neurons in DCN (type I/III units) is not yet clear. Type II units in decerebrate cats were studied using a two-tone paradigm to map inhibitory responses to tones and using noisebands of varying width to study the inhibitory processes evoked by broadband stimuli. Iontophoresis of bicuculline and strychnine and comparisons of two-tone responses between type II units and auditory nerve fibers were used to differentiate inhibitory processes occurring near the cell from two-tone suppression in the cochlea. For type II units, a significant inhibitory region is always seen with two-tone stimuli; the bandwidth of this region corresponds roughly to the previously reported excitatory bandwidth of onset-C neurons. Bandwidth widening experiments with noisebands show a monotonic decline in response as the bandwidth increases; these data are interpreted as revealing strong inhibitory inputs with properties more like onset-C neurons than any other response type in the lower auditory system. Consistent with these properties, iontophoresis of inhibitory antagonists produces a large increase in discharge rate to broadband noise, making tone and noise responses nearly equal.

Responses of auditory-cortex neurons to structural features of natural sounds.

Sound-processing strategies that use the highly non-random structure of natural sounds may confer evolutionary advantage to many species. Auditory processing of natural sounds has been studied almost exclusively in the context of species-specific vocalizations, although these form only a small part of the acoustic biotope. To study the relationships between properties of natural soundscapes and neuronal processing mechanisms in the auditory system, we analysed sound from a range of different environments. Here we show that for many non-animal sounds and background mixtures of animal sounds, energy in different frequency bands is coherently modulated. Co-modulation of different frequency bands in background noise facilitates the detection of tones in noise by humans, a phenomenon known as co-modulation masking release (CMR). We show that co-modulation also improves the ability of auditory-cortex neurons to detect tones in noise, and we propose that this property of auditory neurons may underlie behavioural CMR. This correspondence may represent an adaptation of the auditory system for the use of an attribute of natural sounds to facilitate real-world processing tasks.

Physiology of MPTP tremor.

Rhesus and vervet monkeys respond differently to treatment with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine hydrochloride neurotoxin (MPTP). Both species develop akinesia, rigidity, and severe postural instability. However, rhesus monkeys only develop infrequent, short episodes of high-frequency tremor, whereas vervet monkeys have many prolonged episodes of low-frequency tremor. After MPTP treatment, the spiking activity of many pallidal neurons became oscillatory and highly correlated. Oscillatory autocorrelation functions were dominated by lower frequencies, cross-correlograms by higher frequencies. The phase shift distribution of the oscillatory cross-correlograms of pallidal cells in MPTP-treated vervet monkey were clustered around 0 phase shift, unlike the oscillatory correlograms in the MPTP-treated rhesus monkey, which were widely distributed between 0 degrees and 180 degrees. Analysis of the instantaneous phase differences between tremors of two limbs in the MPTP monkeys and human parkinsonian patients showed short periods of tremor synchronization. We thus concluded that the rhesus and the vervet models of MPTP-induced parkinsonism may represent the tremulous and nontremulous variants of human parkinsonism. We suggest that the tremor phenomena of Parkinson's disease (PD) are related to the emergence of synchronous neuronal oscillations in the basal ganglia. Finally, the oscillating neuronal assemblies in the pallidum of tremulous parkinsonian primates are more stable (in time and in space) than those of parkinsonian primates without overt tremor.

Linear and nonlinear spectral integration in type IV neurons of the dorsal cochlear nucleus. I. Regions of linear interaction.

The principal neurons of the dorsal cochlear nucleus have complex response properties, many of which are classified as type IV. These units integrate energy in the acoustic signal in a nonlinear fashion; for example, at high sound levels the response to a noise of narrow bandwidth and to a band-reject filtered noise with a spectral notch of the same bandwidth may both be inhibitory. However, the sum of these two stimuli, which is broadband noise (BBN), generally gives an excitatory response. In other situations, linear interactions among stimulus components are observed. In this paper, three regimes of approximate linearity were identified. First, best-frequency (BF) tones and equal-energy narrow noisebands centered at BF evoke almost the same response, which is consistent with a stage of linear filtering followed by a nonlinearity that generates the rate responses of the neuron. Second, for sounds close to threshold (10-15 dB re threshold), energy over the full bandwidth of the unit is integrated linearly. Within this regime, responses to the narrow noiseband and the spectral notch mentioned above do sum to equal the response to BBN. Finally, two noisebands centered at different frequencies, such that their sum is a notch in a broad band of noise, sum linearly at low sound levels; the degree of linearity improves as the separation between the noisebands increases. The results are interpreted in terms of a model of type IV response generation containing two inhibitory interneurons: type II units, which are active for narrowband stimuli, including tones, and the wideband inhibitor, which is active for broadband stimuli. In most cases, the onset of nonlinearity occurs for stimuli that significantly activate the type II inhibitory interneuron.

Linear and nonlinear spectral integration in type IV neurons of the dorsal cochlear nucleus. II. Predicting responses with the use of nonlinear models.

Two nonlinear modeling methods were used to characterize the input/output relationships of type IV units, which are one principal cell type in the dorsal cochlear nucleus (DCN). In both cases, the goal was to derive predictive models, i.e., models that could predict the responses to other stimuli. In one method, frequency integration was estimated from response maps derived from single tones and simultaneous pairs of tones presented over a range of frequencies. This model combined linear integration of energy across frequency and nonlinear interactions of energy at different frequencies. The model was used to predict responses to noisebands with varying width and center frequency. In almost all cases, predictions using two-tone interactions were better than linear predictions based on single-tone responses only. In about half the cases, reasonable quantitative fits were achieved. The fits were best for noisebands with narrow bandwidth and low sound levels. In the second nonlinear method, the spectrotemporal receptive field (STRF) was derived from responses to broadband stimuli. The STRF could account for some qualitative features of the responses to broad noisebands and spectral notches embedded in broad noisebands. Quantitatively, however, the STRFs failed to predict the responses of type IV units even to simple broadband noise stimuli. For narrowband stimuli, the STRF failed to predict even qualitative features (such as excitatory and inhibitory frequency bands). The responses of DCN type IV units presumably result from interactions of two inhibitory sources, a strong one that is preferentially activated by narrowband stimuli and a weaker one that is preferentially activated by broadband stimuli. The results presented here suggest that the STRF measures effects related to the broadband inhibition, whereas two-tone interactions measure mostly effects related to narrowband inhibition. This explains why models based on two-tone interactions predict the responses to narrow noisebands much better then models based on STRFs. It is concluded that a minimal stimulus set for characterizing type IV units must contain both broadband and narrowband stimuli, because each stimulus class by itself activates only partially the integration mechanisms that shape the responses of type IV units. Similar conclusions are expected to hold in other parts of the auditory system: when characterizing a complex auditory unit, it is necessary to use a range of stimuli to ensure that all integration mechanisms are activated.

Why do cats need a dorsal cochlear nucleus?

The dorsal cochlear nucleus (DCN), one of the three major divisions of the cochlear nucleus (CN), has a complex internal structure, multiple inputs (some of them non-auditory), and multiple output pathways. Response properties of DCN units are accordingly complex. The principal cells of the DCN have type IV response characteristics, characterized by relatively high levels of spontaneous activity and inhibition by high level best frequency (BF) tones. We showed previously that type IV units are inhibited by two separate inhibitory mechanisms, one of them sensitive to narrow band stimuli and the other to wide band stimuli. One result of the wide band inhibition of type IV units is their sensitivity to spectral notches in the region of their BF - stimuli with such notches inhibit type IV units. The source of the narrow band inhibition is an interneuron in the DCN which has type II response characteristics - it does not have spontaneous activity and is strongly activated by BF tones. The neurons giving rise to type II responses are presumably vertical cells, which also project to other divisions of the CN. From anatomical studies, it is known that type IV units are also inhibited by a third system, which carries non-auditory information; movements of the pinna inhibit type IV units through this system. We hypothesize that type IV units signal important events to the auditory system by being inhibited. Such events are either auditory, e.g. spectral maxima and minima, or non-auditory, such as the somatosensory inputs from the pinnae. We hypothesize that the projection of type II units to the ventral cochlear nucleus (VCN) plays a role in reducing the effects of spectral notches introduced by the pinnae in the core auditory pathway. We conclude that although the DCN lies close to the auditory periphery, it already performs sophisticated tasks of auditory processing.

'Dynamics of neuronal interactions' cannot be explained by 'neuronal transients'.

In a recent paper, Vaadia et al. demonstrated that patterns of firing correlation between single neurons in the cortex of behaving monkeys can be modified within a fraction of a second. These changes occur in relation to sensory stimuli and behavioral events, and even without modulations of the neurons' firing rates. These findings call for a revision of prevailing models of neural coding that solely rely on single neuron firing rates. In a defense of these models, Friston put forward an alternative explanation, proposing that the observed correlation dynamics emerge solely from co-modulations of the firing rates of each of the neurons, while the strength of their interaction remains constant. To test this possibility we re-examined the data, adopting Friston's 'neuronal transients' model, and the associated equations and procedures. We found that, to explain the dynamic correlation between a pair of neurons, the alternative interpretation requires that each neuron's response to a single stimulus is composed of a relatively large number of independent components, which co-vary with their counterparts in the companion neuron. This large number of components and their shapes lead us to conclude that, although in principle possible, the neuronal transients model: (i) does not provide a simpler explanation of the experimental results; and (ii) cannot explain these results without itself deviating significantly from most rate code models.

Somatosensory effects on neurons in dorsal cochlear nucleus.

1. Single units and evoked potentials were recorded in dorsal cochlear nucleus (DCN) in response to electrical stimulation of the somatosensory dorsal column and spinal trigeminal nuclei (together called MSN for medullary somatosensory nuclei) and for tactile somatosensory stimuli. Recordings were from paralyzed decerebrate cats. 2. DCN principal cells (type IV units) were strongly inhibited by electrical stimulation (single 50-microA bipolar pulse) in MSN or by somatosensory stimulation. Units recorded in the fusiform cell and deep layers of DCN were inhibited, suggesting that the inhibition affects both types of principal cells (i.e., both fusiform and giant cells). 3. Interneurons (type II units) that inhibit principal cells were only weakly inhibited by electrical stimulation and were never excited, demonstrating that the inhibitory effect on principal cells does not pass through the type II circuit. In the vicinity of the DCN/PVCN (posteroventral cochlear nucleus) boundary, units were encountered that were excited by electrical stimulation in MSN; some of these neurons responded to sound, and some did not. Their response properties are consistent with the hypothesis that they are deep-layer inhibitory interneurons conveying somatosensory information to the DCN. 4. Analysis of the evoked potentials produced by electrical stimulation in MSN suggests that the somatosensory inputs activate the granule cell system of the DCN molecular layer. A model based on previous work by Klee and Rall was used to show that the distribution of evoked potentials in DCN can be explained as resulting from radial currents produced in the DCN molecular and fusiform-cell layers by synchronous activation of granule cells inputs to fusiform and cartwheel cells. Current-source density analysis of the evoked potentials is consistent with this model. Thus molecular layer interneurons (cartwheel and stellate cells) are a second possible source of inhibition to principal cells. 5. With lower stimulus levels (20 microA) and pulse-pair stimuli (50- to 100-ms interstimulus interval), three components of the inhibitory response can be recognized in both fusiform cell layer and deep layer type IV units: a short-latency inhibition that begins before the start of the evoked potential; a longer-latency inhibition whose timing corresponds to the evoked potential; and an excitatory component that occurs on the rising phase of the evoked potential. The excitatory component is usually overwhelmed by the inhibitory components and could be derived from granule cell inputs; the long-latency inhibitory component could be derived from cartwheel cells or the hypothesized deep-layer inhibitory interneurons.(ABSTRACT TRUNCATED AT 400 WORDS)

Two separate inhibitory mechanisms shape the responses of dorsal cochlear nucleus type IV units to narrowband and wideband stimuli.

1. The principal cells of the dorsal cochlear nucleus (DCN) are mostly inhibited by best frequency (BF) tones but are mostly excited by broadband noise (BBN), producing the so-called type IV response characteristic. The narrowband inhibitory responses can be explained by the inhibitory influence of interneurons with type II response characteristics. However, it is not clear that all the details of the type IV responses can be accounted for by this neural circuit. In particular, many type IV units are inhibited by band-reject noise (notch noise); type II units tend to be only weakly excited by these stimuli, if at all. In this paper we study the relationships between the narrowband, inhibitory and the wideband, excitatory regimens of the type IV responses and present the case for the existence of a second inhibitory source in DCN, called the wideband inhibitor (WBI) below. 2. Type IV units were studied using pure tones, noise bands arithmetically centered on BF, notch noise centered on BF, and BBN. We measured the rate-level function (response rate as function of stimulus level) for each stimulus. This paper is based on the responses of 28 type IV units. 3. Evidence for low-threshold inhibitory input to type IV units is derived from analysis of rate-level functions at sound levels just above threshold. Notch noise stimuli of the appropriate notch width produce inhibition at threshold in this regime. When BBN is presented, this inhibition appears to summate with excitation produced by energy in the band of noise centered on BF, resulting in BBN rate-level functions with decreased slope and maximum firing rate. A range of slopes and maximal firing rates is observed, but these variables are strongly correlated and they are negatively correlated with the strength of the inhibition produced by notch noise; this result supports the conclusion that a single inhibitory source is responsible for these effects. 4. By contrast, there is a weak (nonsignificant) positive correlation between the strength of the inhibitory effect of notch noise and the slope/maximal firing rate in response to narrowband stimuli, including BF tones. The contrast between this positive nonsignificant correlation and the significant negative correlation mentioned above suggests that more than one inhibitory effect operates: specifically, the type II input is responsible for inhibition by narrowband stimuli and a different inhibitory source, the WBI, produces inhibition by notch stimuli. 5. Several lines of evidence are given to show that type II units cannot produce the inhibition seen with notch noise stimuli.(ABSTRACT TRUNCATED AT 400 WORDS)

Population responses to multifrequency sounds in the cat auditory cortex: one- and two-parameter families of sounds.

Population responses to multi-frequency sounds were recorded in primary auditory cortex of anesthetized cats. The sounds consisted of single-tone stimuli; two-tone stimuli; and nine-tone stimuli, with the tones evenly spaced on a linear frequency scale. The stimuli were presented through a sealed, calibrated sound delivery system. Single units, cluster activity (CA) and the short-time mean absolute value of the envelope of the neural signal (MABS) were recorded extracellularly from six microelectrodes simultaneously. The CA and MABS were interpreted as measures of the activity of large populations of neurons, in contrast with the single unit activity which is presumably recorded from single neurons. The responses of the MABS signal to simple stimuli were generally similar to those of the CA, but were more stable statistically. Thus, the MABS is better suited for studying the activity of populations of neurons. The responses to tones near the best frequency were strongly influenced by a second tone, even when the second tone was outside the single-tone response area. These influences could be both facilitatory and suppressory. They could not be predicted from the responses to single tones. The responses to the nine-tone stimuli could be explained qualitatively by the responses to the two-tone stimuli. It is concluded that the population responses in primary auditory cortex are shaped by the contributions of the individual frequencies appearing in the stimulus and by the interactions between pairs of frequencies. Interactions between stimulus components are therefore a necessary component of any attempt to explain the processing of complex sounds in the auditory cortex. They may play a role in a global representation of the stimulus spectrum in the primary auditory cortex. The presence of higher-order interactions cannot be excluded by the results presented here.

Population responses to multifrequency sounds in the cat auditory cortex: four-tone complexes.

Population responses to two-tone and four-tone sounds were recorded in primary auditory cortex of anesthetized cats. The stimuli were delivered through a sealed, calibrated sound delivery system. The envelope of the neural signal (short time mean absolute value, MABS) was recorded extracellularly from six microelectrodes simultaneously. A new method was developed to describe the responses to the four-tone complexes. The responses were represented as sums of contributions of different orders. The first order contributions described the effect of the single frequencies appearing in the stimulus. The second order contributions described the modulatory effect of the pairs of frequencies. Higher order contributions could in principle be computed. This paper concentrates on the mean onset responses. The extent to which the first and second order contributions described the onset responses was assessed in two ways. First, the actual responses to two-tone stimuli were compared with those predicted using the contributions computed from the four-tone stimuli. Second, the residual variance in the responses, after the subtraction of the first and second order contributions, was computed and compared with the variability in the responses to repetitions of the same stimulus. The first type of analysis showed good quantitative agreement between the predicted and the measured two-tone responses. The second type of analysis showed that the first and second order contributions were often sufficient to predict the responses to four-tone stimuli up to the level of the variability in the responses to repetitions of a single stimulus. In conjunction with the results of the companion paper (Nelken et al., 1994a) it is concluded that the onset responses to multifrequency sounds are shaped mainly by the single frequency content of the sound and by two-tone interactions, and that higher order interactions contribute much less to the responses. It follows that single-tone effects and two-tone interactions are necessary and sufficient to explain the mean population onset responses to the four-tone stimuli. More information can be coded in the temporal evolution of the responses.

In search of the best stimulus: an optimization procedure for finding efficient stimuli in the cat auditory cortex.

Units in the auditory cortex of cats respond to a large variety of stimuli: pure tones, AM- and FM-modulated signals, clicks, wideband noise, natural sounds, and more. However, no single family of sounds was found to be optimal (in the sense that oriented lines are optimal in the visual cortex). The search for optimal complex sounds is hard because of the high dimensionality of the space of interesting sounds. In an effort to overcome this problem, an automatic search procedure for finding efficient stimuli in high-dimensional sound spaces was developed. This procedure chooses the stimuli to be presented according to the responses to past stimuli, trying to increase the strength of the response. The results of applying this method to recordings of population activity in the primary auditory cortex of cats are described. The search was applied to single tones, two-tone stimuli, four-tone stimuli and to a two-dimensional subset of nine-tone stimuli, parametrized by the center frequency and the fixed difference between adjacent frequencies. The method was able to find efficient stimuli, and its performance improved with the dimension of the sound spaces. Efficient stimuli, found in different optimization runs using population activity recorded from the same electrode, often shared similar frequencies and pairs of frequencies, and tended to evoke similar levels of activity. This result indicates that a global analysis of the location of spectral peaks is performed at the level of the auditory cortex.

Neural interactions in the frontal cortex of a behaving monkey: signs of dependence on stimulus context and behavioral state.

In order to gain an understanding of the processes taking place within and between neuronal assemblies, we made simultaneous recordings of spike trains from groups of up to 11 neurons in the frontal cortex of a rhesus monkey, that was trained to perform a sensorimotor behavioral task. We report here on preliminary results from correlation analysis of these neuronal activities, with special emphasis on signs of behaviorally induced modifications of neural interaction, possibly due to rapid modulations of discharge synchronization among the neurons. Our findings suggest that different functional groups of neurons may co-exist within each small volume of cortex, and that neurons may be dynamically recruited into such a group to fulfil a specific function.