SearcharxivSearch

arXiv subjects

Ines Samengo

Publications and source records attributed to Ines Samengo.

13 recordsLinked to original sources

The role of the observer in goal-directed behavior

In goal-directed behavior, a large number of possible initial states end up in the pursued goal. The accompanying information loss implies that goal-oriented behavior is in one-to-one correspondence with an open subsystem whose entropy decreases in time. Yet ultimately, the laws of physics are reversible, so systems capable of yielding goal-directed behavior must transfer the information about initial conditions to other degrees of freedom outside the boundaries of the agent. To operate steadily, they must consume ordered degrees of freedom provided as input, and be dispensed of disordered outputs that act as wastes from the point of view of the aimed objective. Here I argue that a physical system may or may not display goal-directed behavior depending on what exactly is defined as the agent. The borders of the agent must be carefully tailored so as to entail the appropriate information balance sheet. In this game, observers play the role of tailors: They design agents by setting the limits of the system of interest. Brain-guided subjects perform this creative observation task naturally, implying that the observation of goal-oriented behavior is a goal-oriented behavior in itself. Minds evolved to cut out pieces of reality and endow them with intentionality, because ascribing intentionality is an efficient way of modeling the world, and making predictions. One most remarkable agent of whom we have indisputable evidence of its goal-pursuing attitude is the self. Notably, this agent is simultaneously the subject and the object of observation.

q-bio.NC

Derivation of human chromatic discrimination ability from an information-theoretical notion of distance in color space

The accuracy with which humans can detect small chromatic differences varies throughout color space. For example, we are far more precise when discriminating two similar orange stimuli than two similar green stimuli. In order for two colors to be perceived as different, the neurons representing chromatic information must respond differently, and the difference must be larger than the trial-to-trial variability of the response to each separate color. Photoreceptors constitute the first stage in the processing of color information; many more stages are required before humans can consciously report whether two stimuli are perceived as chromatically distinguishable or not. Therefore, although photoreceptor absorption curves are expected to influence the accuracy of conscious discriminability, there is no reason to believe that they should suffice to explain it. Here we develop information-theoretical tools based on the Fisher metric that demonstrate that photoreceptor absorption properties explain ~87% of the variance of human color discrimination ability, as tested by previous behavioral experiments. In the context of this theory, the bottleneck in chromatic information processing is determined by photoreceptor absorption characteristics. Subsequent encoding stages modify only marginally the chromatic discriminability at the photoreceptor level.

q-bio.NC

Bursts generate a non-reducible spike pattern code

At the single-neuron level, precisely timed spikes can either constitute firing-rate codes or spike-pattern codes that utilize the relative timing between consecutive spikes. There has been little experimental support for the hypothesis that such temporal patterns contribute substantially to information transmission. By using grasshopper auditory receptors as a model system, we show that correlations between spikes can be used to represent behaviorally relevant stimuli. The correlations reflect the inner structure of the spike train: a succession of burst-like patterns. We demonstrate that bursts with different spike counts encode different stimulus features, such that about 20% of the transmitted information corresponds to discriminating between different features, and the remaining 80% is used to allocate these features in time. In this spike-pattern code, the what and the when of the stimuli are encoded in the duration of each burst and the time of burst onset, respectively. Given the ubiquity of burst firing, we expect similar findings also for other neural systems.

q-bio.NC

The information transmitted by spike patterns in single neurons

Spike patterns have been reported to encode sensory information in several brain areas. Here we assess the role of specific patterns in the neural code, by comparing the amount of information transmitted with different choices of the readout neural alphabet. This allows us to rank several alternative alphabets depending on the amount of information that can be extracted from them. One can thereby identify the specific patterns that constitute the most prominent ingredients of the code. We finally discuss the interplay of categorical and temporal information in the amount of synergy or redundancy in the neural code.

q-bio.QM

The firing statistics of Poisson neuron models driven by slow stimuli

The coding properties of cells with different types of receptive fields have been studied for decades. ON-type neurons fire in response to positive fluctuations of the time-dependent stimulus, whereas OFF cells are driven by negative stimulus segments. Biphasic cells, in turn, are selective to up/down or down/up stimulus upstrokes. In this paper, we explore the way in which different receptive fields affect the firing statistics of Poisson neuron models, when driven with slow stimuli. We find analytical expressions for the time-dependent peri-stimulus time histogram and the inter-spike interval distribution in terms of the incoming signal. Our results enable us to understand the interplay between the intrinsic and extrinsic factors that regulate the statistics of spike trains. The former depend on biophysical neural properties, whereas the latter hinge on the temporal characteristics of the input signal.

q-bio.NC

Quasi-static approximation of the interspike interval distribution of neurons driven by time-dependent inputs

Variability in neural responses is an ubiquitous phenomenon in neurons, usually modeled with stochastic differential equations. In particular, stochastic integrate-and-fire models are widely used to simplify theoretical studies. The statistical properties of the generated spikes depend on the stimulating input current. Given that real sensory neurons are driven by time-dependent signals, here we study how the inter-spike interval distribution of integrate-and-fire neurons depends on the evolution of the stimulus, in a quasi-static limit. We obtain a closed-form expression for this distribution, and we compare it to the one obtained with numerical simulations for several time-dependent currents. For slow inputs, the quasi-static distribution provides a very good description of the data. The results obtained for the integrate-and-fire model can be extended to other non-autonomous stochastic systems where the first passage time problem has an explicit solution.

q-bio.NC

Burst firing is a neural code in an insect auditory system

Various classes of neurons alternate between high-frequency discharges and silent intervals. This phenomenon is called burst firing. To analyze burst activity in an insect system, grasshopper auditory receptor neurons were recorded in vivo for several distinct stimulus types. The experimental data show that both burst probability and burst characteristics are strongly influenced by temporal modulations of the acoustic stimulus. The tendency to burst, hence, is not only determined by cell-intrinsic processes, but also by their interaction with the stimulus time course. We study this interaction quantitatively and observe that bursts containing a certain number of spikes occur shortly after stimulus deflections of specific intensity and duration. Our findings suggest a sparse neural code where information about the stimulus is represented by the number of spikes per burst, irrespective of the detailed interspike-interval structure within a burst. This compact representation cannot be interpreted as a firing-rate code. An information-theoretical analysis reveals that the number of spikes per burst reliably conveys information about the amplitude and duration of sound transients, whereas their time of occurrence is reflected by the burst onset time. The investigated neurons encode almost half of the total transmitted information in burst activity.

q-bio.NC

Dynamics of learning in coupled oscillators tutored with delayed reinforcements

In this work we analyze the solutions of a simple system of coupled phase oscillators in which the connectivity is learned dynamically. The model is inspired in the process of learning of birdsong by oscine birds. An oscillator acts as the generator of a basic rhythm, and drives slave oscillators which are responsible for different motor actions. The driving signal arrives to each driven oscillator through two different pathways. One of them is a "direct" pathway. The other one is a "reinforcement" pathway, through which the signal arrives delayed. The coupling coefficients between the driving oscillator and the slave ones evolve in time following a Hebbian-like rule. We discuss the conditions under which a driven oscillator is capable of learning to lock to the driver. The resulting phase difference and connectivity is a function of the delay of the reinforcement. Around some specific delays, the system is capable to generate dramatic changes in the phase difference between the driver and the driven systems. We discuss the dynamical mechanism responsible for this effect, and possible applications of this learning scheme.

q-bio.NC

Independent neurons representing a finite set of stimuli: dependence of the mutual information on the number of units sampled

We study the capacity with which a system of independent neuron-like units represents a given set of stimuli. We assume that each neuron provides a fixed amount of information, and that the information provided by different neurons has a random overlap. We derive analytically the dependence of the mutual information between the set of stimuli and the neural responses on the number of units sampled. For a large set of stimuli, the mutual information rises linearly with the number of neurons, and later saturates exponentially at its maximum value.

cond-mat.dis-nn

Representational capacity of a set of independent neurons

The capacity with which a system of independent neuron-like units represents a given set of stimuli is studied by calculating the mutual information between the stimuli and the neural responses. Both discrete noiseless and continuous noisy neurons are analyzed. In both cases, the information grows monotonically with the number of neurons considered. Under the assumption that neurons are independent, the mutual information rises linearly from zero, and approaches exponentially its maximum value. We find the dependence of the initial slope on the number of stimuli and on the sparseness of the representation.

cond-mat.dis-nn

Estimating probabilities from experimental frequencies

Estimating the probability distribution 'q' governing the behaviour of a certain variable by sampling its value a finite number of times most typically involves an error. Successive measurements allow the construction of a histogram, or frequency count 'f', of each of the possible outcomes. In this work, the probability that the true distribution be 'q', given that the frequency count 'f' was sampled, is studied. Such a probability may be written as a Gibbs distribution. A thermodynamic potential, which allows an easy evaluation of the mean Kullback-Leibler divergence between the true and measured distribution, is defined. For a large number of samples, the expectation value of any function of 'q' is expanded in powers of the inverse number of samples. As an example, the moments, the entropy and the mutual information are analyzed.

cond-mat.stat-mech

Measuring information spatial densities

A novel definition of the stimulus-specific information is presented, which is particularly useful when the stimuli constitute a continuous and metric set, as for example, position in space. The approach allows one to build the spatial information distribution of a given neural response. The method is applied to the investigation of putative differences in the coding of position in hippocampus and lateral septum.

cond-mat.dis-nn

Information loss in an optimal maximum likelihood decoding

The mutual information between a set of stimuli and the elicited neural responses is compared to the corresponding decoded information. The decoding procedure is presented as an artificial distortion of the joint probabilities between stimuli and responses. The information loss is quantified. Whenever the probabilities are only slightly distorted, the information loss is shown to be quadratic in the distortion

physics.bio-ph