Spike timing, synchronization and information processing on the sensory side of the central nervous system.

To what extent is the variability of the neuronal responses compatible with the use of spike timing for sensory information processing by the central nervous system? In reviewing the state of the art of this question, I first analyze the characteristics of this variability with its three elements: synaptic noise, impact of ongoing activity and possible fluctuations in evoked responses. I then review the recent literature on the various sensory modalities: somato-sensory, olfactory, gustatory and visual and auditory processing. I emphasize that the conditions in which precise timing, at the millisecond level, is usually obtained, are conditions that usually require dynamic stimulation or sharp changes in the stimuli. By contrast, situations in which stimulation not belonging to the temporal domain is temporally encoded lead to much coarser temporal coding; although in both cases, neural networks transmit the signals with similarly high precision. Synchronization among neurons is an important tool in information processing in both cases but again seems to act either at millisecond or tens of millisecond levels. Information theory applied to both situations confirms that the average rate of information transmission is much higher in dynamic than in static situations. These facts suggest that channels of precise temporal encoding may exist in the brain but imply populations of neurons working in a yet to be discovered way.

[A network capable of reading the temporal codes in the olfactory discrimination of insects]. / Un réseau capable de lire les codes temporels dans la discrimination olfactive chez l'insecte.

Recent work on the insect olfactory system has shown that its mushroom bodies (one of its major components) are involved in the fine discrimination of odours and that the temporal organisation of spike discharges plays a fundamental role. We propose here a model of a network that is able to decode the temporal patterns which characterise an odour. This model has three fundamental properties that seem to exist in all mushroom bodies of insects studied so far: a) long lasting inhibitions with rebounds, able to facilitate delayed spike generation; b) synaptic plasticity, which allows the network to learn to recognise temporal patterns; c) above all a large interconnection, which allows this network to recognise intervals of various duration. This model thus appears suited to identify combinations of temporal patterns in the dendrites of Kenyon cells (the principal cells in the calyces of the mushroom bodies). Moreover, the mushroom bodies integrate multimodal inputs, suggesting that the detection of temporal patterns may be extended to the detection of a complex environment, combining in particular olfactive and visual inputs.

Repeating triplets of spikes and oscillations in the mitral cell discharges of freely breathing rats.

The olfactory bulb responses to odours display evident temporal organization, both in the form of high-frequency oscillations and precisely replicating triplets of spikes. In this study, the frequency of replicating triplets in a sample of 118 individual responses from 45 cells was compared with that in simulations of non-homogeneous Poisson processes, constructed from the experimental post-stimulus time histograms (PSTHs). In a large majority of the records, replicating triplets (to a precision of 0.5 ms) are found to be more numerous in the physiological records; in some of them, they are approximately 10 times more abundant. An excess of precisely replicating triplets is also found in records where no oscillations are apparent in the autocorrelograms. Triplet replication thus seems a more robust phenomenon than transient oscillation. Not unlike fast oscillations observed in other preparations, replicating triplets produced by a given mitral cell are generally observed only during a restricted period of time of the respiratory cycle (at least in the case of the responses under olfactory stimulation). No relation was found, however, between the nature and strength of the olfactory stimulus and the frequency of replicating patterns. In the absence of olfactory stimulation, some mitral cell discharges also contain more replicating triplets than the non-homogeneous Poisson simulations. Thus, replicating triplets in single-cell discharges seem to play only an indirect role in the coding of olfactory information at the mitral cell output level.

Stochastic nature of precisely timed spike patterns in visual system neuronal responses.

It is not clear how information related to cognitive or psychological processes is carried by or represented in the responses of single neurons. One provocative proposal is that precisely timed spike patterns play a role in carrying such information. This would require that these spike patterns have the potential for carrying information that would not be available from other measures such as spike count or latency. We examined exactly timed (1-ms precision) triplets and quadruplets of spikes in the stimulus-elicited responses of lateral geniculate nucleus (LGN) and primary visual cortex (V1) neurons of the awake fixating rhesus monkey. Large numbers of these precisely timed spike patterns were found. Information theoretical analysis showed that the precisely timed spike patterns carried only information already available from spike count, suggesting that the number of precisely timed spike patterns was related to firing rate. We therefore examined statistical models relating precisely timed spike patterns to response strength. Previous statistical models use observed properties of neuronal responses such as the peristimulus time histogram, interspike interval, and/or spike count distributions to constrain the parameters of the model. We examined a new stochastic model, which unlike previous models included all three of these constraints and unlike previous models predicted the numbers and types of observed precisely timed spike patterns. This shows that the precise temporal structures of stimulus-elicited responses in LGN and V1 can occur by chance. We show that any deviation of the spike count distribution, no matter how small, from a Poisson distribution necessarily changes the number of precisely timed spike patterns expected in neural responses. Overall the results indicate that the fine temporal structure of responses can only be interpreted once all the coarse temporal statistics of neural responses have been taken into account.

Intrinsic and extrinsic neuronal mechanisms in temporal coding: a further look at neuronal oscillations.

Many studies in recent years have been devoted to the detection of fast oscillations in the Central Nervous System (CNS), interpreting them as synchronizing devices. We should, however, refrain from associating too closely the two concepts of synchronization and oscillation. Whereas synchronization is a relatively well-defined concept, by contrast oscillation of a population of neurones in the CNS looks loosely defined, in the sense that both its frequency sharpness and the duration of the oscillatory episodes vary widely from case to case. Also, the functions of oscillations in the brain are multiple are not confined to synchronization. The paradigmatic instantiation of oscillation in physics is given by the harmonic oscillator, a device particularly suited to tell the time, as in clocks. We will thus examine first the case of oscillations or cycling discharges of neurones, which provide a clock or impose a "tempo" for various kinds of information processing. Neuronal oscillators are rarely just clocks clicking at a fixed frequency. Instead, their frequency is often adjustable and controllable, as in the example of the "chattering cells" discovered in the superficial layers of the visual cortex. Moreover, adjustable frequency oscillators are suitable for use in "phase locked loops" (PLL) networks, a device that can convert time coding to frequency coding; such PLL units have been found in the somatosensory cortex of guinea pigs. Finally, are oscillations stricto sensu necessary to induce synchronization in the discharges of downstream neurones? We know that this is not the case, at least not for local populations of neurones. As a contribution to this question, we propose that repeating patterns in neuronal discharges production may be looked at as one such alternative solution in relation to the processing of information. We review here the case of precisely repeating triplets, detected in the discharges of olfactory mitral cells of a freely breathing rat under odor stimulation.

The significance of precisely replicating patterns in mammalian CNS spike trains.

Neuronal spike trains from both single and multi-unit recordings often contain patterns such as doublets and triplets of spikes that precisely replicate themselves at a later time. The presence of such precisely replicating patterns can still be detected when the tolerance on interval replication is shortened to a fraction of a millisecond. In this context we examine here data taken from various parts of the central nervous systems of anesthetized rats, cats and monkeys. The relative abundance of replicating triplets varies from centre to centre, and is nearly always significantly greater than obtained in Monte-Carlo simulations of either a Poisson-like process or a renewal process having the same interspike interval distribution as the neuronal data. However, a remarkable exception is found in the activity of retinal ganglion cells. Significant deviations were found in the primary visual cortex and, even more so, in the lateral geniculate body and the mitral cells of the olfactory bulb. Using a fixed tolerance for the replication of intervals (0.5 ms) it is usually observed that replicating patterns are produced in excess (with respect to renewal process models) mostly in low firing rate episodes (< or = 100 Hz). However, using a tolerance that varies in direct proportion to the mean interval (i.e. as the reciprocal of the firing rate), one generally observes that replicating triplets occur with higher than expected frequency in comparable proportions at all firing rates. This observation suggests the existence of a scale invariance principle in these phenomena with respect to certain neuronal codes. In order to decrease the influence of the estimated neuronal firing rate on the results of the comparisons, we computed also the ratio NT2/ND3, of the number of replicating triplets to the number of doublets replicating three times [Lestienne R. (1994) Proc. Soc. Neurosci. 20, 22; Lestienne R. (1996) Biol. Cybern. 74, 55-61], using both a fixed or a variable tolerance. In spike trains obeying a Poisson process, NT2/ND3 ratios should be nearly independent of the frequency, especially when using a variable tolerance. These studies supported previous results: significant deviations from the models are found in all the spike trains examined, except in the case of retinal ganglion cells, and the most significant deviations are found in recordings from the lateral geniculate nucleus and the mitral cells of the olfactory bulb. Removing spikes that belong to bursts having large "Poisson surprise" values [Legéndy C. R. and Salcman M. (1985) J. Neurophysiol. 53, 926-939] (except the very first spike of the burst) significantly decreases NT2/ND3 ratios in the record from the lateral geniculate nucleus, suggesting that in this case bursty episodes greatly contribute to the production of replicating patterns, but such a removal does not affect results from the piriform record. Finally, in both the lateral geniculate nucleus and in the mitral cells of the olfactory bulb records, perturbing the timing of spikes by applying to interspike intervals small jitters of uniform probability density with amplitude up to 3 ms, very significantly decrease NT2/ND3 ratios in these centres, but does not change much the NT2/ND3 ratios in other neuronal recordings. Implications of these findings for a possible role of precisely replicating patterns in temporal coding of neuronal information is discussed, as well as possible mechanisms for their production.

Slow oscillations as a probe of the dynamics of the locus coeruleus-frontal cortex interaction in anesthetized rats.

Multiunit or single unit activity recorded simultaneously from frontal cortex (FC) and locus coeruleus (LC) under ketamine anesthesia revealed that both regions show slow oscillatory activity, together or separately. If, however, both regions are engaged in this oscillatory activity, there is a systematic relationship between their phases with peak LC firing always following FC firing by 200-400 ms. This was confirmed by cross-correlational analyses, which indicated that the two structures temporarily form a resonant system. The FC-LC resonant state is, however, loose enough to remain open to other intrinsic or extrinsic influences, keeping the measured frequencies of oscillations at each site slightly different, as demonstrated by a detailed analysis of the autocorrelograms. An injection of lidocaine at the frontal cortex site, while sharply reducing the prefrontal activity to essentially zero, leads to an increase of the LC activity and to a modification of the shape of the LC autocorrelogram, but does not change appreciably the phase relationship between the activity in the two structures during the diminishing activity in FC.

Determination of the precision of spike timing in the visual cortex of anaesthetised cats.

Neuronal cortical spike trains contain precisely replicating patterns whose presence cannot be accounted for by change production. A comparison of the number of triplets of spikes present two times with the number of doublets replicated three times in the same window duration gives a frequency-insensitive measure of this type of fine temporal organisation. By varying the tolerance with which such precisely replicating patterns are detected, one can evaluate the accuracy of spike timing in spike trains. In the sample of data here analysed, it was found that replicating patterns were best seen in the precision range 0.4-1.4 ms (a result evidently at variance with a simple 'integrate and fire' model of neurons). Surprisingly, the fine temporal structure of spike trains thus evidenced was stronger at relatively low firing rate discharges and was present in both the 'spontaneous' and 'evoked' responses.

[Cortical modulations of the fine temporal structures of impulse trains in the dorsal lateral geniculate body in cats]. / Modulations corticales de la structure temporelle fine des trains d'impulsions dans le corps genouillé latéral dorsal du Chat.

In response to a visual stimulation, "replicated triplets" of impulses may appear in many spike trains recorded from the cat dorsal lateral geniculate nucleus (dLGN). The number and the temporal structure of these triplets depend upon the general organization of the geniculate impulse trains. In this study, we show that a pharmacological blockade of the corticothalamic activity, obtained through microinjection of GABA into area 17, affects the replicated triplet production and leads to an increase in the dispersal of their structure. These results suggest that the corticothalamic pathway closely influences the fine temporal organization of the geniculate messages.

Temporal correlations in modulated evoked responses in the visual cortical cells of the cat.

We analysed evoked responses recorded from 97 cells in the visual cortex of 4 adult cats and 8 kittens, stimulated by a drifting sinusoidal grating. A Fourier analysis of the responses allowed us to select 30 cells showing a clear modulating response (relative modulation index greater than 1). The 162 records from these selected cells were scanned to detect precise temporal correlations in the form of replicating triplets and associated "ghost" doublets. Temporal correlations of this nature were observed in these cells. They are about 10 times more abundant in adult cats than in kittens, and mostly observed in infragranular cortical layer cells. The possible role of these precise temporal patterns in information processing in the brain is examined, as well as the relation between this type of temporal correlation with coherent oscillations and principal components waveforms.

Presence of ghost doublets of coded neuronal patterns: relation to synaptic memory storage.

Recent evidence demonstrates that controlled visual stimuli cause the generation, in the primary visual cortex of rhesus monkey cells, of large numbers of very precisely replicating copies of complex patterns of discharge consisting of three or more spikes, the patterns of which presumedly code for specific qualities of the stimuli presented. We present evidence that the copies of precisely replicating triplets of spikes, generally not exceeding 100 ms in duration, occur in close time proximity to many copies of highly precise "ghost" doublets. These doublets are defined as patterns consisting of two pulses, with precise separations in time, specifically those that would be generated if any one of the pulses making up a given replicating triplet were missing. In striking contrast, nonreplicating triplets (also present in these records)--that is, triplets made up of intervals that are not present in replicating triplets--are not accompanied by such ghost doublets. The persistence (memory) of capacity to produce such ghost doublets decays according to two independent kinetic rules. The first of these results in the disappearance of such doublets within about 0.1 s as measured by two independent methods, whereas the second disappears only after several minutes or longer. These results provide strong evidence consistent with the notion that at least some parts of the brain transmit, store representations of, and retrieve qualitative information through the use of a code consisting of specific patterns of nerve discharges in time.

From physical to biological time.

Time is a primitive (i.e. fundamental) notion, and the various concepts** that have been so far derived from this notion in various scientific domains do not cover all facets of it. Time in mechanics, either classical, quantal or relativistic, is devoid of directionality, the "arrow", i.e. of irreversibility: these physical theories are fundamentally time reversible. Thermodynamics, however, does involve irreversibility, but only as an empirical observation, rather than as a fundamental law of nature (entropy decreasing processes are said to be very improbable, they are not said to be forbidden). In contrast with physics, the arrow of time is of a tremendous importance and effect in biology. For this reason here I will propose a notion of time that--contrary to the claim of several current epistemological schools--is both primitive and oriented. Time flow and irreversibility are indeed at the heart of phenomena of the generation and growth of biological order (in developing organisms), of phenomena of maintenance of organisms in their healthy adult age (it is suggested that the production and coordination of temporal cycles are as important and perhaps more important in understanding this maintenance than the usually emphasized phenomenon of homeostasis), and finally of phenomena of senescence, with their ultimate issue: death. In all these fields, life obeys--does not negate--the thermodynamic law of increase of entropy, as the development of irreversible processes thermodynamics allows us to understand it. Many biologists use the concept of entropy in a somewhat restricted, and sometimes misleading way, namely as a measure of disorder. But the relationship between entropy and disorder is more subtle than mere equivalence. In order to clarify these ideas, a most precise relationship between entropy and order, using the physical concept of phase space, is expounded and illustrated. Application of the results of thermodynamics of irreversible processes to living beings requires a jump in complexity, the wideness of which is acknowledged: possible specific effects of this jump (linked for instance to the high number of hierarchical levels and/or to the role of randomness in the organization of the lower levels) are mentioned.4+ Some current models of aging, using thermodynamical analogies, are examined and discussed . Finally, it is pointed out that the concepts of time so far examined do not include the notion of the "present", which is so obviously at the heart of our psychological, internal and subjective notion of time.(ABSTRACT TRUNCATED AT 400 WORDS)

On the thermodynamical and biological interpretation of the Gompertzian mortality rate distribution.

"Thermodynamical" foundations of a Gompertzian representation of mortality rate distribution vs. age are reviewed. The two fundamental assumptions of the original model (as developed by Strehler and Mildvan) are: (1) challenges that threaten the lives of organisms of a given species have an exponential distribution in harmfulness; and (2) vitality ("energetic" reserve to be used to counteract the challenges and restore proper function of the organism) declines linearly with age. It is proposed that the external environment should not only be characterized by a "temperature", but also by a "pressure" (related to the average time between successive hits). While recent progress of health sciences have essentially lowered the "temperature" factor, future progress might also lower the "pressure" factor. The effect of this would be to provide only a slight extension of observed longevity in humans. Internal cause(s) of ageing that lead to death are not specified in the model (except that they should be compatible with the linear decline in vitality). It is shown that death cannot be attributed to a slowing down of the recovery machinery that restores the organism's state following a challenge or a disease. A mechanism of this kind would instead lead to a gamma- (rather than a Gompertzian) distribution of ages at death, at great ages. Whatever the modalities of the challenges (they are, of course, not necessarily of a literally energetic nature), the model is shown to assume that death is linked to single, large amplitude challenges, rather than to the conjunction of independent, small amplitude damages. The concept of programmed longevity is proposed and integrated into the model. In this new model, Gompertzian distributions are characterized by the two parameters alpha (slope) and L (longevity) rather than by the two traditional parameters alpha and R0 (mortality rate at birth). This new presentation is more parsimonious than the original one, in that only alpha (not L) is temperature dependent. Models with fixed longevity automatically display a negative correlation between ln R0 and alpha, as was noted by Strehler and Mildvan. There exists a definite lag of time (of 23-29 years) between longevity and the most probable age at death. Assuming that the human species has a maximum programmed longevity of 120 years, this implies that the progress of health sciences will allow the present survival curve to evolve, not towards a rectangular shape as previously believed, but rather to a given limiting curve such as is depicted.

Differences between monkey visual cortex cells in triplet and ghost doublet informational symbols relationships.

On the basis of the recent discovery that precisely replicating triplets of impulses present in All-Interval histograms of spike trains generated by visual cortex cells of Rhesus monkeys are surrounded by multiple copies of "ghost doublets" of such triplets, we have examined and compared in detail, the spike trains generated by four complex cells in the striate cortex of curarized monkeys with respect to: (1) The number of precisely replicating triplet patterns embedded in trains of discharges generated in response to specific Hubel-Wiesel stimulation; (2) The effect of time separating the occurrence of such replicating triplets on the number and time distribution of their ghost doublets; (3) The effect of decreasing the precision criterion for the detection of replicating (parent) triplets (from the standard 0.14 ms criterion to 0.5 ms) on the relationships between triplets and their ghosts and (4) The comparison of the distributions in time of ghost doublets around the first and second copies of triplets when the time intervals separating them were greater than or less than 0.5 s. We found that the precision of replication of triplets varies somewhat from one cell to another, and that ghosts doublets are more copiously associated with replicating triplets emitted near in time to each other than with triplets emitted after larger time intervals, except in the case of one cell. In order to assess the statistical significance of our findings, we systematically shuffled the order of occurrence of intervals in every burst of all the records of one of the studied cells and repeated the analysis. Both the number of replicating triplets and of associated ghost doublets is significantly depressed (but not totally obliterated) by the above shuffling procedure. Finally, further implications based on a model of neural information transmission in the form of temporal correlations between spikes are discussed.

[Temporal correlations and coding of information in the visual cortex of the cat]. / Corrélations temporelles et codage de l'information dans le cortex visuel du chat.

We have observed repeated patterns in evoked spike trains recorded from the primary visual cortex of the cat. These patterns are called "triplets" and "ghost doublets". Triplets are groups of three pulses, that may or may not be adjacent to one other, the mutual intervals of which are replicated in one other group of three spikes with a precision higher than 0.15 ms. Ghost doublets are doublets of pulses whose interval replicates, with the above precision, one of the intervals of the repeated triplets and are also present in the record. In one of the 9 recorded cells, in which pulses were clearly emitted in bursts in phase with the drifting of the sinusoidal grating used as a stimulus, we could show that local temporal correlations in the form of replicating triplets and ghost doublets correspond very precisely to the temporal phase of the grating: the study of the distance between triplets, or between triplets and ghost doublets, gives a remarkably precise value of the time frequency of the grating.

Time structure and stimulus dependence of precisely replicating patterns present in monkey cortical neuronal spike trains.

Evidence is presented on the parameters that affect the occurrence of precisely replicating patterns of neural discharge present as 'hidden' patterns in individual neuronal discharge trains of the visual cortical cells of the rhesus monkey in response to precisely controlled stimuli described in our previous publication. Using the All-Interval analytical paradigm we demonstrate: (1) that precisely replicating patterns are present in numbers that cannot be generated through continuous, smoothly varying probability distributions of interspike intervals; (2) that the records contain very large numbers of precisely replicating patterns--doublets, triplets, quadruplets, quintuplets and hextuplets of pulses; (3) that triplet-antitriplet pairs and symmetrical quadruplets are also present in improbable numbers; (4) that different stimuli generate different triplets; (5) and that the first order decay constant of capacity to generate specific precise patterns is a direct function of the number of events making up the patterns and thus that a temporary memory of the occurrence of a pattern exists following the presentation of a stimulus. It is concluded that such patterns of pulses are almost certainly coded symbols related to visual information; that such symbols are sufficiently precise in their replication to permit them to be decoded through spatial summation mechanisms and finally that the ability to generate and the capacity to store such symbols are probably present in the brain as related and coordinated complexes of specific facilitated synapses. Some properties of a proposed model for the production and decoding of such patterns are presented and discussed as are specific mechanisms through which neural networks may implement such functions. Finally, existing and further experimental tests of the mechanisms proposed are outlined.

Evidence on precise time-coded symbols and memory of patterns in monkey cortical neuronal spike trains.

High-resolution examination of pulse sequences generated by single visual cortex cells of the rhesus monkey in response to precisely controlled visual stimuli has disclosed (i) that the outputs of such neurons contain highly improbable (P less than 10(-7) numbers of identical triplets of precisely repeating pulse patterns; (ii) that the precision of such matches is better than 1/6000th of a second; (iii) that there is a similarly improbable high number of precisely matching pairs of triplets and anti-triplets, about half of which are present in symmetrical quadruplets of the form A-B-A, and that precisely replicating quadruplets and quintuplets are similarly generated in improbably large numbers; (iv) that identical triplets occur highly preferentially during immediately succeeding presentation of the same stimulus to the eye; and (v) that identical triplet (and doublet) patterns occur much more frequently in the responses of the same nerve when the eye receives identical or similar stimuli in different experiments than when dissimilar stimuli are applied. From these findings it is concluded (i) that the high precision of pattern replication required for triplets of pulses in time to serve to encode specific inputs and to permit their decoding through spatial summation is met (observations i-iii); (ii) that stimulus-specific triplets symbolize components of responses to specific stimuli (observation iv); (iii) that a temporary memory store of previous responses exists (observations iv and v); and (iv) that the mammalian brain uses precise patterns of discharges in time to represent and store specific data, rather than statistical qualities associated with pulse trains to symbolize qualitative stimulus components.