Correlations and the encoding of information in the nervous system

Is the information transmitted by an ensemble of neurons determined solely by the number of spikes fired by each cell, or do correlations in the emission of action potentials also play a significant role? We derive a simple formula which enables this question to be answered rigorously for short time-scales. The formula quantifies the corrections to the instantaneous information rate which result from correlations in spike emission between pairs of neurons. The mutual information that the ensemble of neurons conveys about external stimuli can thus be broken down into firing rate and correlation components. This analysis provides fundamental constraints upon the nature of information coding, showing that over short time-scales correlations cannot dominate information representation, that stimulus-independent correlations may lead to synergy (where the neurons together convey more information than they would if they were considered independently), but that only certain combinations of the different sources of correlation result in significant synergy rather than in redundancy or in negligible effects. This analysis leads to a new quantification procedure which is directly applicable to simultaneous multiple neuron recordings.     

Changing reference frames during the encoding of tactile events

The mindless act of swatting a mosquito on the hand poses a remarkable challenge for the brain. Given that the primary somatosensory cortex maps skin location independently of arm posture [1, 2], the brain must realign tactile coordinates in order to locate the origin of the stimuli in extrapersonal space. Previous studies have highlighted the behavioral relevance of such an external mapping of touch, which results from combining somatosensory input with proprioceptive and visual cues about body posture [3-7]. However, despite the widely held assumption about the existence of this remapping process from somatotopic to external space and various findings indirectly suggesting its consequences [8-11], a demonstration of its changing time course and nature was lacking. We examined the temporal course of this multisensory interaction and its implications for tactile awareness in humans using a crossmodal cueing paradigm [12, 13]. What we show is that before tactile events are referred to external locations [12-15], a fleeting, unconscious image of the tactile sensation abiding to a somatotopic frame of reference rules performance. We propose that this early somatotopic "glimpse" arises from the initial feed-forward sweep of neural activity to the primary somatosensory cortex, whereas the later externally-based, conscious experience reflects the activity of a somatosensory network involving recurrent connections from association areas.     

Neural mechanisms of absolute tactile localization in monkeys

Macaca nemestrina monkeys were trained to indicate the location of suprathreshold tactile stimuli delivered to the glabrous skin of either foot. The testing paradigm involved self-initiated trials (a bar press), followed by 10-Hz stimulation at one of six locations (e.g., on the distal phalanx of the second toe on the left foot), providing the opportunity for the animal to press one of six buttons located on a facing panel. The buttons were positioned on a picture of a monkey's feet at locations corresponding to the skin loci that were stimulated on different trials. If the animal first pressed the button corresponding to the position stimulated, liquid reward was delivered; responses to any other button terminated stimulation without reward, requiring initiation of another trial for the opportunity to receive reinforcement. The localization errors for normal monkeys were reliably greater along the mediolateral dimension of the foot than they were proximodistally. For example, stimulation of the tip of toe 4 elicited responses to the button at the tip of toe 2 on 25% of the trials, as compared with only 10% errors between the tip of toe 4 and the pad at the base of toe 4. Following unilateral interruption of the dorsal spinal columns at an upper thoracic level, the capacity for absolute tactile localization was unchanged over months of testing. The greater localization accuracy along the proximodistal axis of the foot remained after dorsal column transection. In order to evaluate neural substrates of localization by monkeys, single-neuron receptive field (RF) sizes and distributions within the first somatosensory (SI) cortex were examined to determine the overlap or separation of the representations of different points on glabrous skin. The sample of neurons that provided the RF data was obtained in previous investigations of unanesthetized, neuromuscularly blocked Macaca fascicularis monkeys. Analysis of RF overlap revealed that greater than 50% of cytoarchitectural area 1 units that responded to stimulation of one digit tip also responded to another digit or to the pad at the base of a digit. These large RFs seem poorly suited to subserve a high degree of spatial localization and are compatible with the frequent localization errors by the monkeys in the behavioral experiments. However, the area 1 RF data do not explain the tendency of these animals to exhibit better localization accuracy along the proximodistal axis than along the mediolateral axis of the volar foot.(ABSTRACT TRUNCATED AT 250 WORDS)     

Transformation of the afferent tactile signal into a motor command in the cat motor cortex

Activity of 112 neurons of the precruciate motor cortex in cats was studied during a forelimb placing reaction to tactile stimulation of its distal parts. The latent period of response of the limb to tactile stimulation was: for flexors of the elbow (biceps brachii) 30–40 msec, for the earliest reponses of cortical motor neurons about 20 msec. The biceps response was observed 5–10 msec after the end of stimulation of the cortex with a series of pulses lasting 25 msec. Two types of excitatory responses of the neurons were identified: responses of sensory type observed to each tactile stimulation of the limb and independent of the presence or absence of motion, and responses of motor type, which developed parallel with the motor response of the limb and were not observed in the absence of motion. The minimal latent period of the responses of motor type was equal to the latent period of the sensory responses to tactile stimulation (20±10 msec). Stimulation of the cortex through the recording microelectrode at the site of derivation of unit activity, which increased during active flexion of the forelimb at the elbow (11 stimuli at intervals of 2.5 msec, current not exceeding 25 µA), in 70% of cases evoked an electrical response in the flexor muscle of the elbow.M. V. Lomonosov Moscow State University. Translated from Neirofiziologiya, Vol. 9, No. 2, pp. 115–123, March–April, 1977.     

Spinal control of afferent temperature information in the cat

The effect of picrotoxin (at sub-convulsant doses, i.v.) on the background and invoked spike activity of cat dorsal horn internerons was investigated while thermal and mechanical influences were at work on the skin receptor areas. It was found that information on skin temperature is modulated at the presynaptic level and largely mediated by competitive interaction between large and small diameter fibers.Institute of Physiology, Academy of Sciences of the Kazakh SSR, Alma Ata. Translated from Neirofiziologiya, Vol. 18, No. 6, pp. 773–779, November–December, 1986.     

Neural activity and survival in the developing nervous system

Recent evidence suggests that blockade of normal excitation in the immature nervous system may have profound effects on neuronal survival during the period of natural cell death. Cell loss following depression of electrical activity in the central nervous system (CNS) may explain the neuropsychiatric deficits in humans exposed to alcohol or other CNS depressants during development. Thus, understanding the role of electrical activity in the survival of young neurons is an important goal of modern basic and clinical neuroscience. Here we review the evidence from in vivo and in vitro model systems that electrical activity participates in promoting neuronal survival. We discuss the potential role of moderate elevations of intracellular calcium in promoting survival, and we address the possible ways in which activity and conventional trophic factors may interact.     

Rigorous and extended application of information theory to the afferent visual system of the cat

1. Intracellular recording were obtained from P-cells of the LGN of the cat. The impulse trains of a single presynaptic retinal ganglion cell and the postsynaptic P-cell were separated by band-pass-filtering and subsequent amplitude discrimination.
2. The rates of information and transinformation for the visual channel from the eye to a ganglion cell and to the connected P-cell were calculated. Input signals to the channel were trains of light flashes of different rate, luminance and spatial distribution.
3. Transinformation was calculated without restrictive assumptions for the code.
4. The transient behaviour of the system in response to a flash was fully considered for information calculations. Additionally, it was ensured that the state of the (adaptive) channel was considered correctly.
5. Information theory was applied in an extended way. The time courses of information transfer were calculated for various flash stimuli and compared with each other.

     

Neural mechanisms of tactile motion integration in somatosensory cortex

How are local motion signals integrated to form a global motion percept? We investigate the neural mechanisms of tactile motion integration by presenting tactile gratings and plaids to the fingertips of monkeys, using the tactile analogue of a visual monitor and recording the responses evoked in somatosensory cortical neurons. The perceived directions of the gratings and plaids are measured in parallel psychophysical experiments. We identify a population of somatosensory neurons that exhibit integration properties comparable to those induced by analogous visual stimuli in area MT and find that these neural responses account for the perceived direction of the stimuli across all stimulus conditions tested. The preferred direction of the neurons and the perceived direction of the stimuli can be predicted from the weighted average of the directions of the individual stimulus features, highlighting that the somatosensory system implements a vector average mechanism to compute tactile motion direction that bears striking similarities to its visual counterpart.     

Radial distance determination in the rat vibrissal system and the effects of Weber's law

Rats rhythmically tap and brush their vibrissae (whiskers) against objects to tactually explore the environment. To extract a complex feature such as the contour of an object, the rat must at least implicitly estimate radial object distance, that is, the distance from the base of the vibrissa to the point of object contact. Radial object distance cannot be directly measured, however, because there are no mechanoreceptors along the vibrissa. Instead, the mechanical signals generated by the vibrissa's interaction with the environment must be transmitted to mechanoreceptors near the vibrissa base. The first part of this paper surveys the different mechanical methods by which the rat could determine radial object distance. Two novel methods are highlighted: one based on measurement of bending moment and axial force at the vibrissa base, and a second based on measurement of how far the vibrissa rotates beyond initial contact. The second part of the paper discusses the application of Weber's law to two methods for radial distance determination. In both cases, Weber's law predicts that the rat will have greatest sensing resolution close to the vibrissa tip. These predictions could be tested with behavioural experiments that measure the perceptual acuity of the rat.     

Characteristics of the formation of temperature information in afferent units

The impulse activity of single afferent fibers of the dorsal roots of the cat spinal cord is studied for local mechanical, heat, and cold influences on the skin receptor fields. A probability analysis of the impulse flux suggests that a change occurs in the distributions of the intervals between impulses in accordance with the stimuli presented, regardless of the variations in the mean frequency of impulsation. It is hypothesized that the afferent fibers acquire polyfunctional properties on account of their multichannel information.Institute of Physiology, Kazakhstan, Academy of Sciences, Alma-Ata. Translated from Neirofiziologiya, Vol. 24, No. 5, pp. 582–591, September–October, 1992.     

Nonlinear neuronal mode analysis of action potential encoding in the cockroach tactile spine neuron

 Neuronal mode analysis is a recently developed technique for modelling the behavior of nonlinear systems whose outputs consist of action potentials. The system is modelled as a set of parallel linear filters, or modes, which feed into a multi-input threshold. The characteristics of the principal modes and the multi-input threshold device can be derived from Laguerre function expansions of the computed first- and second-order Volterra kernels when the system is stimulated with a randomly varying input. Neuronal mode analysis was used to model the encoder properties of the cockroach tactile spine neuron, a nonlinear, rapidly adapting, sensory neuron with reliable behavior. The analysis found two principal modes, one rapid and excitatory, the other slower and inhibitory. The two modes have analogies to two of the pathways in a block-structured model of the encoder that was developed from previous physiological investigations of the neuron. These results support the block-structured model and offer a new approach to identifying the components responsible for the nonlinear dynamic properties of this neuronal encoder. Received: 30 December 1994/Accepted in revised form: 25 April 1995     

Hair cycle-dependent changes of alkaline phosphatase activity in the mesenchyme and epithelium in mouse vibrissal follicles

Alkaline phosphatase (ALP) activity was detected in the restricted mesenchymal and epithelial regions in mouse vibrissal follicles. Its localization and strength dramatically changed during the hair cycle. Activity in the dermal papilla (DP) was moderate in very early anagen, reached a maximal level in early anagen, decreased at the proximal region of DP after mid anagen, and was kept at a low level during catagen. The bulbar dermal sheath showed intense ALP activity only in early anagen. Although most bulbar epithelium did not show ALP activity, germinative epidermal cells that were adjacent to the ALP-negative DP cells became ALP-positive in mid anagen and rearranged in a single layer so as to encapsulate the DP in mid catagen. During catagen, the outermost layer of bulbar epithelium became ALP-positive, which could be follicular epithelial precursors migrating from the bulge. Before the initiation of hair formation, ALP activity in the bulbar epithelium rapidly decreased and that in DP increased. These dynamic changes of ALP expression might be related to DP's functions in hair induction and also to reconstruction of the bulbar structure during the hair cycle.