|
ScienceWeek
NEUROBIOLOGY: ON THE NEURAL CODE
The following points are made by M.R. DeWeese and A. Zador (Nature 2006 439:320):
1) Our perception of the outside world relies on the transformation of physical signals (such as light and sound) into a pattern of neural impulses, or spikes. These spikes are then transmitted to higher brain regions, where they are further transformed into other patterns of sensory spikes, and ultimately into the motor spikes that mediate behavior. What is the relationship (the "neural code") between these neural responses and the sensory signals they represent? Are there general principles underlying the neural code?
2) The "efficient-coding hypothesis"[1] proposes that sensory neurons are adapted to the statistical properties of sensory signals to which neurons are exposed. New work[2,3] invokes this principle to predict how neurons encode natural auditory and visual stimuli, as opposed to the artificial stimuli often used in experiments. Smith and Lewicki[2] develop an algorithm to find an efficient representation of natural sounds and speech, and show that this theoretically predicted representation matches that observed experimentally in the auditory nerve of cats. Sharpee et al[3] show that cortical neurons adapt over seconds or minutes during the course of an experiment to maximize the information they provide about the stimulus. Together, the two reports show how the efficient-coding hypothesis can help to make sense of properties of the neural code on both evolutionary and behavioral timescales.
3) In the cochlea, sound is encoded into spikes, which are transmitted along the auditory nerve to higher stations in the auditory system. Auditory nerve fibers each respond to a narrow range of sound frequencies, with the range generally increasing with the median frequency. The response of each auditory nerve fiber can therefore be modelled as a (nonlinear) "filter" that removes frequencies outside a particular range. Why do the auditory nerve filters have the particular form they do? Smith and Lewicki[2] reasoned that if the auditory code is indeed "efficient", then they should be able to predict the form of the auditory filter bank by finding the sparsest code; that is, the one that requires the least activity.
4) To obtain this prediction, Smith and Lewicki[2] first expressed the efficient-coding hypothesis as an algorithm whose input is an ensemble of sounds, and whose output is a sparse encoding for transmitting or representing this ensemble. The algorithm discovers that the sparsest encoding of sounds is into brief events suggestive of spikes, the precise timing of which conveys much of the information. The sparsest code depends on the ensemble of sounds to be encoded; a code that is most efficient for one set of sounds is not necessarily most efficient for another.[4,5]
References:
1. Barlow, H. B. in Information Processing in the Nervous System (ed. Leibovic, K. N.) 209-230 (Springer, New York, 1969)
2. Smith, E. C. & Lewicki, M. S. Nature 439, 978-982 (2006)
3. Sharpee, T. O. et al. Nature 439, 936-942 (2006)
4. Chklovskii, D. B. , Schikorski, T. & Stevens, C. F. Neuron 34, 341-347 (2002)
5. Laughlin, S. B. , de Ruyter van Steveninck, R. R. & Anderson, J. C. Nature Neurosci. 1, 36-41 (1998)
Nature http://www.nature.com/nature
--------------------------------
Related Material:
ADAPTATION AND NEURAL CODES
The following points are made by A.L. Fairhall et al (Nature 2001 412:787):
1) In 1954, it was suggested by F. Attneave that neural codes may constitute efficient representations of the sensory world. Within the framework of information theory, efficient coding requires a matching of the coding strategy to the statistics of the input signals. Recent work demonstrates that the sequences of action potentials from single neurons provide an efficient representation of complex dynamic inputs, that adaptation to the distribution of inputs can occur in real time, and that the form of the adaptation can serve to maximize information transmission.
2) However, adaptation involves compromises. An adaptive code is inherently ambiguous: the meaning of a spike or a pattern of spikes depends on context, and resolution of this ambiguity requires that the system additionally encode information about the context itself. In a dynamic environment, the context changes in time and there is a trade-off between tracking rapid changes and optimizing the code for the current context.
3) The authors report they examined the dynamics of adaptation to statistics in motion-sensitive cells in the visual system of a fly (Calliphora vicina), and they find that different aspects of adaptation occur on timescales that range from tens of milliseconds to several minutes. The speed of adaptation to a new input distribution must be limited by the need to collect statistics. Adaptation of the neural input/output relation to optimize information transmission approaches this theoretical maximum speed. This rapid adaptation of the input/output relation leaves the longer timescales in the response dynamics as a nearly independent channel for information, resolving potential ambiguities of an adaptive code.
Nature http://www.nature.com/nature
--------------------------------
Related Material:
INTERACTIVE PROCESSING OF SENSORY INPUT AND MOTOR OUTPUT IN THE HUMAN HIPPOCAMPUS.
The following points are made by C. D. Teschea and J.Karhua (J. Cognitive Neurosci. 1999 11:424):
1) A widely held view of brain organization is that the motor system is activated only after the immediate sensory scene is fully elaborated in sensory networks. However, recent studies of visuomotor integration suggest that the motor system may be intimately involved in the detection of relevant features of the sensory scene. The final stages of sensory processing involved in memory encoding and novelty detection occur in hippocampal structures. The authors suggest that large-scale cognitive networks also may recruit additional resources from the hippocampus during sensorimotor integration in humans.
2) Hippocampus and cortico-hippocampal networks have been traditionally associated with memory encoding. The foremost function of reciprocal cortico-hippocampal connections is believed to be the rapid and accurate exchange of information between sensory cortical areas and hippocampal structures for the encoding of memory traces and possibly for the immediate comparison of novel input with stored traces. Hippocampal structures thus participate in the selection of pertinent information that needs to be held "online" during the temporal interval required for a decision or for the performance of an operation, that is, to working memory. In everyday life, these functions are exercised during, and often form an inseparable part of, goal-directed motor activity.
3) The timing of the attentional enhancement of hippocampal neuronal population responses has been evaluated in humans by depth electrode recordings performed on neurological patients and by magnetoencephalographic (MEG) studies performed on normal subjects. Event-related potentials (ERPs) and fields (ERFs) have been observed in the medial temporal lobe to attended infrequent deviants embedded in trains of standard stimuli. Results include task-dependent ERPs and ERFs at peak latencies of 300 to 600 msec following auditory, visual, and somatosensory oddballs. However, the dynamic interactions between motor networks and hippocampal structures are poorly known.
4) In summary: Recent studies of visuomotor integration suggest that the motor system may be intimately involved in the detection of salient features of the sensory scene. The final stages of sensory processing occur in hippocampal structures. The authors report they measured human neuromagnetic responses during motor reaction to an auditory cue embedded in high-speed multimodal stimulation. The authors suggest their results demonstrate that large-scale cognitive networks may recruit additional resources from the hippocampus during sensorimotor integration. Hippocampal activity from 300 msec before to 200 msec after cued movements was enhanced significantly over that observed during self-paced movements. The dominant hippocampal activity appeared equally synchronized to both sensory input and motor output, consistent with timing by an intrinsic mechanism, possibly provided by ongoing theta oscillations.
ScienceWeek http://scienceweek.com
|