Do neurons process information by relative intervals in spike trains?

We suggest the possibility that neurons process information in terms of the relative duration of clusters of adjacent and successive inter-action potential intervals (“bytes” of intervals). If this concept is plausible, as is supported by research from several laboratories which have specifically addressed this posibility, one should be able to see evidence for such patterning in the published illustrations from studies in which this concept was not considered. We present some of this evidence here, along with some illustrations from the original publications. Byte patterns are evident in these examples, even though they often went unrecognized by authors and readers alike. It is true that interval patterns are not obvious in all published illustrations of spike trains, and we suggest that this can be explained by one or more of the following: (1) some neurons may operate with an interval-pattern code while others do not, (2) a given neuron may use an interval-pattern code only under certain conditions, and (3) even when such a code exists, it may be difficult to detect for identifiable technical reasons. Therefore, we believe that the relative-interval-pattern concept is a valid scientific hypothesis which merits specific testing of its validity and range of applicability.     

Do neurons process information by relative intervals in spike trains?

We suggest the possibility that neurons process information in terms of the relative duration of clusters of adjacent and successive inter-action potential intervals (“bytes” of intervals). If this concept is plausible, as is supported by research from several laboratories which have specifically addressed this posibility, one should be able to see evidence for such patterning in the published illustrations from studies in which this concept was not considered. We present some of this evidence here, along with some illustrations from the original publications. Byte patterns are evident in these examples, even though they often went unrecognized by authors and readers alike. It is true that interval patterns are not obvious in all published illustrations of spike trains, and we suggest that this can be explained by one or more of the following: (1) some neurons may operate with an interval-pattern code while others do not, (2) a given neuron may use an interval-pattern code only under certain conditions, and (3) even when such a code exists, it may be difficult to detect for identifiable technical reasons. Therefore, we believe that the relative-interval-pattern concept is a valid scientific hypothesis which merits specific testing of its validity and range of applicability.     

Operant control of precentral neurons: Control of modal interspike intervals

The objects of these expirements were: (a) to determine modal interspike intervals (ISIs) of precentral cells involved in repetitious, gross motor movements; (b) to compare those modal ISIs to the modal ISIs of similar neurons under operant control; and (c) to determine if monkeys could change the modal ISIs of operantly controlled precentral neurons. Data were obtained from 4 monkeys conditioned to produce tonic firing of precentral neurons and one monkey trained to produce repetitious movements of the neck and contralateral limbs. Results are: (a) the modal ISIs from operantly controlled precentral units do not differ significantly from precentral neurons involved in repetitive gross motor movements; and (b) while under operant control, the monkeys cannot modify significantly the modal ISI of the majority of precentral neurons.     

Predominant information transfer from layer III pyramidal neurons to corticospinal neurons

Connections of layer III pyramidal neurons to corticospinal neurons of layer V and corticothalamic neurons of layer VI in the rat primary motor cortex were examined in brain slices by combining intracellular staining with Golgi-like retrograde labeling of corticofugal neurons. Forty layer III pyramidal neurons stained intracellularly were of the regular-spiking type, showed immunoreactivity for glutaminase, and emitted axon collaterals arborizing locally in layers II/III and/or V. Nine of them were reconstructed for morphologic analysis; 15.2% or 3.8% of varicosities of axon collaterals of the reconstructed neurons were apposed to dendrites of corticospinal or corticothalamic neurons, respectively. By confocal laser scanning and electron microscopy, some of these appositions were revealed to make synapses. These findings suggest that corticospinal neurons receive information from the superficial cortical layers four times more frequently than corticothalamic neurons. The connections were further examined by intracellular recording of excitatory postsynaptic potential (EPSP) that were evoked in layer V and layer VI pyramidal neurons by stimulation of layer II/III. EPSPs evoked in layer V pyramidal neurons showed short and constant onset latencies, suggesting their monosynaptic nature. In contrast, most EPSPs evoked in layer VI pyramidal neurons had long onset latencies, showed double-shock facilitation of onset latency, and were largely suppressed by an N-methyl-D-aspartic acid receptor blocker, suggesting that they were polysynaptic. The results suggest that information from the superficial cortical layers is transferred directly and efficiently to corticospinal neurons in layer V and thereby exerts an important influence on cortical motor output. Corticothalamic neurons are, in contrast, considered relatively independent of, or indirectly related to, information processing of the superficial cortical layers.     

Efficient information transfer and anti-Hebbian neural networks

We attempt to maximise mutual information in a neural network where the output signal is corrupted by noise in the transmission process to the next stage. We assume that the average power available for transmission is limited and that the transmission noise is uncorrelated equal-variance Gaussian noise. Using the Lagrange multiplier technique, we construct a utility function that should be maximised to achieve our optimum. This optimum will be achieved when the output signal components are also uncorrelated and of equal variance. For a linear network with lateral inhibitory connections, using an anti-Hebbian algorithm with modified self-inhibitory connections we guarantee to increase our utility function over time and converge to the required optimum, although not by steepest ascent. For a network with inhibitory interneurons, a combined Hebbian/anti-Hebbian algorithm will similarly increase the utility function over time to find the optimum. This network with interneurons suggests that negative feedback may be used in a perceptual system to optimise forward transmission of information.     

Judging time intervals using a model of perceptuo-motor control

Estimating a time interval and temporally coordinating movements in space are fundamental skills, but the relationships between these different forms of timing, and the neural processes that they incur, are not well understood. While different theories have been proposed to account for time perception, time estimation, and the temporal patterns of coordination, there are no general mechanisms which unify these various timing skills. This study considers whether a model of perceptuo-motor timing, the tau(GUIDE), can also describe how certain judgements of elapsed time are made. To evaluate this, an equation for determining interval estimates was derived from the tau(GUIDE) model and tested in a task where participants had to throw a ball and estimate when it would hit the floor. The results showed that in accordance with the model, very accurate judgements could be made without vision (mean timing error -19.24 msec), and the model was a good predictor of skilled participants' estimate timing. It was concluded that since the tau(GUIDE) principle provides temporal information in a generic form, it could be a unitary process that links different forms of timing.     

Reliability and information transmission in spiking neurons.

Spiking neurons encode continuous, time-varying signals in sequences of identical action potentials. Relatively simple algorithms allow one to 'decode' this neural representation of sensory data to estimate the input signals. Decoding experiments provide a quantitative characterization of information transmission and computational reliability under real-time conditions. The results of these studies show that neural coding and computation in several systems approach fundamental physical and informational theoretic limits to performance.     

Efficient transduction of spiral ganglion neurons in vitro by baculovirus vectors

Degeneration of peripheral auditory neurons constitutes one of the main causes of sensorineural hearing loss. Gene delivery to the inner ear is central to the development of gene therapy for hearing impairment. Thus we investigated the effectiveness of baculovirus-derived vectors to transduce spiral ganglion neurons. We found that baculovirus could efficiently transduce spiral ganglion neurons in vitro and that the highest transduced cell rate could be over 75%. The level of transgene expression exhibited viral dose dependence and was enhanced by the addition of butyrate. Thus, baculovirus is a novel and promising tool for gene transfer into the cochlear nervous system, both in studies of the function of foreign genes and in the development of gene-therapy strategies.     

Metathoracic neurons integrating intersegmental sensory information in the locust

This paper describes the morphology and physiology of five types of local interneurons and three types of ascending intersegmental interneurons in the locust metathoracic ganglion that are points of convergence of sensory information from the wings. Four types of spiking local interneurons are members of a population with somata at the ventral midline. They are depolarised by stimulation of a metathoracic wing nerve, suggesting that they encode a sensory representation of this appendage. Some are also depolarised with short latencies following stimulation of a mesothoracic wing nerve, indicating that they collate intersegmental as well as local information. All the local interneurons have branches in the anterior ventral association centre or around the roots of the nerve that carries wing sensory neurons. This distinguishes them from other interneurons in the population. A fifth type of local interneuron that has unusual bilateral branching and is not a member of this population is described for the first time. The ascending interneurons are members of three populations. Neurons of each population have a characteristic pattern of responses to stimulation of the mesothoracic or metathoracic wing nerves, and some respond to tactile stimulation or movements of a hind leg. These latter interneurons thus collate information from both wings and legs. All three types of intersegmental interneurons have branches in the anterior ventral association centre or around the roots of the wing nerve. The responses of the interneurons described here shed new light on both local and intersegmental network function in this model system.     

Graded information extraction by neural-network dynamics with multihysteretic neurons

A major goal in the study of neural networks is to create novel information-processing algorithms inferred from the real brain. Recent neurophysiological evidence of graded persistent activity suggests that the brain possesses neural mechanisms for retrieval of graded information, which could be described by the neural-network dynamics with attractors that are continuously dependent on the initial state. Theoretical studies have also demonstrated that model neurons with a multihysteretic response property can generate robust continuous attractors. Inspired by these lines of evidence, we proposed an algorithm given by the multihysteretic neuron-network dynamics, devised to retrieve graded information specific to a given topic (i.e., context, represented by the initial state). To demonstrate the validity of the proposed algorithm, we examined keyword extraction from documents, which is best fitted for evaluating the appropriateness of retrieval of graded information. The performance of keyword extraction by using our algorithm was significantly high (measured by the average precision of document retrieval, for which the appropriateness of keyword extraction is crucial) compared with standard document-retrieval methods. Moreover, our algorithm exhibited much higher performance than the neural-network dynamics with bistable neurons, which can also produce robust continuous attractors but only represent dichotomous information at the single-cell level. These findings indicate that the capability to manage graded information at the single-cell level was essential for obtaining a high performing algorithm.     

Efficient production of neural stem cells and neurons from embryonic stem cells

We have developed a simple method to efficiently produce a large number of neural stem cells and neurons from mouse embryonic stem (ES) cells. When cultured in astrocyte-conditioned medium (ACM) with mitogens (FGF-2 and EGF) under free-floating conditions, colonies of undifferentiated ES cells give rise to neural stem spheres (NSSs), composed of plentiful neural stem cells. Subsequent culture of the NSSs on an adhesive substrate with mitogens results in the migration of neural stem cells onto the substrate. These cells can be expanded, preserved by freezing, and differentiated into functional neurons. Neural stem cells and neurons provided by this NSS method may be valuable as potential donor cells for neuronal transplantation and also as convenient alternatives to tissue-derived neural cells.     

Variability of interspike intervals of human motor neurons during voluntary muscle contraction and tonic vibration reflex]

Firing of motor units of human soleus, triceps brachii and rectus femoris muscles was studied. Standard deviations of interspike intervals against mean intervals were plotted during voluntary muscle contraction and tonic vibration reflex. There was no significant difference between the results obtained under these conditions.     

Efficient expression of tetracycline-responsive gene after transfection of dentate gyrus neurons in vitro

Gene transfer into neurons both in vivo and in vitro may aid in understanding of gene regulation and function in nerve cells. Especially desirable is ability to control the gene expression. In this study we developed conditions for transfection of hippocampal dentate gyrus neurons in dissociated cultures in vitro by calcium-phosphate method. Furthermore, we describe an effective use of tetracycline responsive gene promoter (Tet-On) system for the controlled and very efficient expression of transfected genes. Under optimal conditions as established in this study, efficiency of transfection of neurons with green fluorescent protein (GFP) driven by constitutive cytomegalovirus (CMV) early promoter reached 2.7%. With tetracycline responsive promoter percentage of GFP-positive neurons raised in the presence of tetracycline analog, doxycycline up to 20%. Application of the Tet-On system resulted in almost 10-fold induction of GFP expression.     

Progressive morphological abnormalities observed in rat spinal motor neurons at extended intervals after axonal regeneration.

It is generally accepted that mammalian spinal motor neurons return to normal after axotomy if their regenerated axons successfully reinnervate appropriate peripheral targets. However, morphological abnormalities, recently observed in spinal motor neurons examined 1 year after nerve crush injury, raise the possibility that delayed perikaryal changes occur after regeneration is complete. In order to distinguish between chronic and progressive alterations in neurons with long-term regenerated axons, rat spinal motor neurons and dorsal root ganglion cells were examined at 5 and 10 months following unilateral sciatic nerve crush. Neurons with regenerated axons were identified by retrograde labelling with horseradish peroxidase. The structural properties of neurons ipsilateral to nerve injury were compared to those of neurons from the spinal cord and dorsal root ganglia on the contralateral side and from age-matched control rats. At 5 months postcrush, the morphology of motor and sensory neurons ipsilateral to injury was comparable to that of control cells. However, several features of the motor neurons with regenerated axons distinguished them from control motor neurons at 10 months postcrush. Mean perikaryal area of ipsilateral spinal motor neurons was larger than the means for control motor neurons (p less than .001). Ipsilateral spinal motor neurons also appeared clustered within the spinal cord and had thicker dendrites. Dorsal root ganglion cells with regenerated axons were slightly larger than control cells at 10 months postcrush but they exhibited no other morphological changes. The present findings indicate that spinal motor neurons are progressively altered after their regenerated axons have reestablished functional synapses with their peripheral targets.     

Contribution of spike timing to the information transmitted by HVC neurons

In many species, neurons with highly selective stimulus-response properties characterize higher order sensory areas and/or sensory motor areas of the CNS. In the songbird nuclei, the responses of HVC (used as a proper name) neurons during playback of the bird's own song (BOS) are probably one of the most striking examples of selectivity for natural stimuli. We examined here to what extent spike-timing carries information about natural and time-reversed versions of the BOS. From a heterogenous population of 107 HVC neurons recorded in long-day or short-day conditions, a standard indicator of stimulus preference based on spike-count (the d' index) indicates that a limited proportion of cells can be classified as selective for the BOS (20% with a |d'| > 1). In contrast, quantifying the information conveyed by spike trains with the metric-space of J.D. Victor & K.P Purpura [(1996) J. Neurophysiol., 76, 1310-1326] indicates that 62% of the cells display significant amounts of transmitted information, among which 77% are 'temporal cells'. 'Temporal cells' correspond to cells transmitting significant amounts of information when spike-timing is considered, whereas no information, or lower amounts of transmitted information, is obtained when only spike-count is considered. Computing a correlation index between spike trains [S. Schreiber et al. (2003) Neurocomputing, 52-54,925-931] revealed that spike-timing reliability is higher for the forward than for the reverse BOS, whatever the day length and the cell type are. Cells classified as selective in terms of spike-counts (d' index) had greater amounts of transmitted information, but cells classified as non-selective (d' < 0.5) can also transmit significant amounts of information. Thus, information theory methods demonstrate that a much larger proportion of neurons than expected based on spike-count only participate in the discrimination between stimuli.     

Intrinsic signal optical imaging of brain function using short stimulus delivery intervals

Intrinsic signal optical imaging (ISOI) can be used to map cortical function and organization. Because its detected signal lasts 10+ s consisting of three phases, trials are typically collected using a long (tens of seconds) stimulus delivery interval (SDI) at the expense of efficiency, even when interested in mapping only the first signal phase (e.g., ISOI initial dip). It is unclear how the activity profile can change when stimuli are delivered at shorter intervals, and whether a short SDI can be implemented to improve efficiency. The goals of the present study are twofold: characterize the ISOI activity profile when multiple stimuli are delivered at 4 s intervals, and determine whether successful mapping can be attained from trials collected using an SDI of 4 s (offering >10× increase in efficiency). Our results indicate that four stimuli delivered 4 s apart evoke an activity profile different from the triphasic signal, consisting of signal dips in a series at the same frequency as the stimuli despite a strong rise in signal prior to the 2nd to 4th stimuli. Visualization of such signal dips is dependent on using a baseline immediately prior to every stimulus. Use of the 4-s SDI is confirmed to successfully map activity with a similar location in peak activity and increased areal extent and peak magnitude compared to using a long SDI. Additional experiments were performed to begin addressing issues such as SDI temporal jittering, response magnitude as a function of SDI duration, and application for successful mapping of cortical function topography.