On the Principle of Specific Coding

In the inner ear (cochlea) the acoustical stimulus is encoded into the ensemble of pulse series occurring in each of the 40,000 nerve fibers of the auditory nerve. The cochlea exhibits, on hydrodynamic grounds, a frequency-to-space transformation with a modest amount of frequency resolution. For sinusoidal stimuli the nerve fibers show a far greater frequency selectivity. The instants at which action potentials (nerve pulses of uniform waveform) may occur in an individual nerve fiber can be predicted from a signal transformation model which contains as its most essential elements a linear filter followed by a triggerable pulse generator. This model explains frequency selectivity and phase locking for sinusoidal stimuli in a satisfactory way, provided the correct parameters are selected in accordance with the specific properties of the nerve fiber under study. Whether such, a model would represent frequency resolution in a more general sense, remains to be seen. As far as the linear circuit, part one of the model, is concerned, application of a cross-correlation technique under stimulation with white noise would yield the filter’s impulse response characteristic. However, in the physiological experiment the output of the filter is not accessible. It has been shown that with a special correlation technique, utilizing the (analog) stimulus signal and the (digital) series of action potentials of a nerve fiber, it is possible to recover the essential properties of the linear filter’s impulse response. Application of this “reverse correlation” technique in experiments on anaesthetized cats has shown that under stimulation with white noise the filter has a very sharp frequency response. This effective frequency response agrees well with the one obtained with sinusoidal signals. That this response is so much sharper than the mechanics of the cochlea would allow for, remains a puzzling, and as yet unexplainable, fact. It is concluded that frequency analysis in the cochlea proceeds as if it were realized by a linear filter and the initiation of nerve pulses is a process that operates quite independently of it. Each one of the nerve fibers of the auditory nerve is apparently excited by a specific portion of the acoustical stimulus’ frequency spectrum; the “resonance frequencies” of the fibers covering the entire range of audible frequencies. This property is referred to as the “principle of specific coding.” The findings bear an interesting relation to properties of the (human) auditory system that have been obtained by psychophysical experiments. From several problem areas one can infer that the manner of signal encoding as described by the principle of specific coding is not exhaustive. It may well be possible that finer details about the excitation pattern of nerve fibers are processed by higher auditory centers.

Encoding and the rea for speech signals

A number of different possible explanations are distinguished for the findings on the right ear advantage (REA) for speech signals varied in their acoustic and phonetic properties. Two experiments are reported, using synthesized semivowels and vowels in monosyllable word frames. Both show REA. The detailed results of both experiments support the idea that a complicated “encoded” relationship between the acoustical stimulus and the response phoneme is a necessary condition for the REA, but that the encoding need only be signalled by an acoustical “trigger feature” in the stimulus; a task requiring a perceptual decoding is not necessary for REA to occur.

An Analysis of the Phonoresponse of Males of the True Katydid, Pterophylla Camellifolia (Fabricius) (Orthoptera: Tettigoniidae)

Abstract1. Males of the true katydid, Pterophylla camellifolia F., produce three kinds of acoustical signals termed calling (which can be subdivided into solo calling and alternating calling), aggressive, and disturbance sounds. Alternating calling and aggressive sounds are specialized phonoresponses consisting of regular, rhythmic alternation of chirps by adjacent males at rates slower than solo calling. The arhythmic disturbance sounds are elicited by handling. 2. The nature of alternation was investigated by analyzing changes in chirp lengths and interval lengths of males responding to chirps of other males and to electronically-produced imitations of their chirps for which rate, duration, and intensity was varied. 3. During alternating calling and aggressive sounds, a katydid's chirp rate is slowed because of delay or inhibition by the chirp of another katydid or imitation chirp. Katydids are refractory to delay until the acoustical stimulus extends to the point at which a katydid would chirp if not delayed. From this point to the point of maximum delay (the alternation period), limited by the length of the stimulus and the solo rate of the katydid, increase in the interval between the stimulus and the previous katydid chirp results in a relatively constant interval between the stimulus and the following katydid chirp. During alternating calling, the chirps of one katydid extend beyond the refractory period of the other katydid so that the intervals between successive chirps remain relatively constant. 4. Acoustical interaction between two katydids consists essentially of: 1) entrainment of each katydid at a slower rate because of inhibition by the acoustical stimulus (chirp of the other katydid), and 2) intermittent "escapes" from entrainment (solos). The katydid with the faster spontaneous rate (the leader) is responsible for almost all solos; response by the katydid with the slower spontaneous rate (the follower) eventually entrains (slows) the leader and alternation is resumed. 5. Excitation is as much an integral part of the alternation phonoresponse as is inhibition. Post-inhibitory excitation is apparently responsible for the following: i) katydids soloing faster following than prior to alternation, 2) increase in chirp length following delay by an acoustical stimulus, 3) katydids chirping only in response to an "inhibitory" acoustical stimulus, and 4) reduction in the difference between the solo (spontaneous) rates of two alternating katydids. 6. Increasing the interval between a katydid's chirp and a succeeding stimulus chirp not only increases the extent of delay, it also increases the probability that the next chirp will be one pulse longer. Shortening of a chirp by one pulse occurs most frequently when a katydid's chirp begins before the end, or within a few hundredths of a second after the end, of the stimulus chirp. Under these conditions, shortening of chirp length occurs whether or not the katydid's chirp rate has been slowed. Acoustically-generated nervous signals must partially inhibit nervous output to the wing muscles even though there may be no effect on chirp rate. 7. An increase in the length of electronic stimulus chirps causes an equivalent decrease in katydid chirp rate. Increase in the period of inhibition also results in an increase in post-inhibitory excitation expressed in the following ways: i) increase in solo rate, 2) shortening of intervals between alternated chirps, and 3) to some extent for some katydids, an increase in frequency of longer chirps. 8. The aggressive sound, which occurs at inter-katydid distances of from one to seven feet (sound intensities of 69 to 80 dB), is characterized by a one- to six-pulse increase in chirp length and up to a 50% decrease in intervals between solo chirps. Increasing the intensity of imitation (electronic) stimulus chirps above a threshold value of approximately 65 dB has little or no effect on a katydid's acoustical response. Increasing the length of electronic stimulus chirps causes either no change or no more than a one-pulse increase in katydid chirp length and no more than a 25% decrease in intervals between solo chirps. It is suggested that non-acoustical stimuli may be responsible for release of the aggressive sound. 9. Most of the acoustical behavior of Pterophylla camellifolia males can be understood simply in terms of afferent inhibition of a spontaneously firing neuron ("acoustic pacemaker") and post-inhibitory rebound; this suggests a neuronal mechanism consisting of relatively few neurons. Utilizing information uncovered from the investigation of central nervous system activity in other species of arthropods, some likely properties of the neuronal mechanism controlling katydid phonoresponding are discussed. 10. The nature of alternation of chirps by other tettigoniids suggests that the neuronal mechanism may be similar in all alternating tettigoniids. Differences in the nature of alternation may be related to differences in the chirp-duration-to-chirp-interval ratios and differences in the expression of inhibitory and post-inhibition excitatory processes. 11. The relatively long chirps and short intervals between chirps of P. camellifolia males result in maximal expression of reciprocal inhibitory-excitatory processes which, in turn, may be responsible for the following: i) more katydids chirping more continuously for longer periods of time, and 2) maximum clarity and redundancy of species-specific signals by maintenance of constant response intervals between chirps.