Frequency Difference Limens Research Articles

Comparative frequency discrimination in primates

Auditory frequency difference limens at 2.0 kHz were measured monaurally in three humans and three Old World monkeys: one Japanese macaque (Macaca fuscata), one vervet (Cercopithecus aethiops), and one De Brazza monkey (Cercopithecus neglectus). A go, no‐go, repetitive‐standard discrimination procedure was employed with positive reinforcement. In addition to species differences, we examined the effects of presence versus absence of feedback after missed stimuli, changes in sensation level (35 vs 45 dB), and presenting frequency increments versus decrements. Results showed that human ΔFs (4–11 Hz) were superior to those of monkeys (25–99 Hz), and all subjects performed better with feedback after misses. In addition, while the best human thresholds tended to occur with frequency increments at higher sensation levels, the best monkey thresholds tended to occur with frequency decrements at lower sensation levels. Of the three monkeys, the Japanese macaque exhibited performance qualitatively and quantitatively...

Auditory signal processing in hearing-impaired persons with residual high-frequency hearing

A number of persons with severe sensorineural hearing losses in the frequency range below 6000 Hz have been shown to have near-normal thresholds for frequencies above 6000 Hz. This paper reports on measures of high-frequency signal processing in seven persons with this type of heating impairment. Measures taken include estimates of difference limens for intensity and frequency, gap detection, rate discrimination, and narrow-band masking functions. Although little normative data exist for comparative purposes, measures of frequency, intensity, and time resolution are within the range of what one would anticipate by extrapolation. Intensity and frequency differences limens, and wideband and narrow-band masking functions also were obtained in the region of impaired hearing {i.e., 3–6 kHz}; narrow-band masking measures and difference limens indicated impaired frequency processing and varied between subjects. Intensity discrimination fell within previously published ranges. [Work supported by the National Institutes of Health.]

Cross‐correlation studies of whole nerve action potential waveforms in chinchillas

A cross correlation was calculated for the compound action potential (CAP) response to a tone burst at a given reference frequency at 10 dB above CAP threshold and the compound action potential response to test signals at adjacent frequencies. The normalized cross‐correlation coefficient at zero time delay was measured as a function of the test frequency with signal level maintained at the 10 dB level above threshold. Constant CAP peak‐to‐peak magnitude and constant sound pressure level responses were also correlated with the response to the reference tone burst. Repeating the experimental procedure at several reference frequencies, the correlation coefficients formed band‐pass functions centered about the respective reference frequencies. The data were examined for a possible relationship between the bandwidths of the correlation coefficient functions and critical bandwidths and/or frequency difference limens. [Supported by the NINCDS.]

Frequency discrimination by infants and adults

Frequency difference thresholds were estimated for infants (5 to 8 months old) using the visually reinforced head‐turn procedure. Stimuli were 500‐ms tone bursts at six frequencies ranging from 250 to 8000 Hz presented monaurally via a TDH‐49 earphone. Tone bursts were presented at 70 dB above adult absolute threshold. An adaptive threshold estimation procedure with a variable step size was used. An adult group was run using the same protocol. The relative difference threshold (Δf/f) was a nonmonotonic function of frequency with a minimum around 1000 Hz for both age groups. However, infants had higher thresholds than adults at frequencies from 250 to 2000 Hz (approximately 2% for infants versus 1% for adults). No age difference in threshold was observed at 4000 or 8000 Hz. Age differences in absolute sensitivity at low frequencies [Trehub et al., J. Exp. Child Psychol. 21, 219–225] are not great enough to account for the rather large difference in discrimination performance in the low‐frequency range. It is suggested that these findings reflect continuing development of the analytic capacity of the auditory system during the first year of life. [Work supported by Virginia Commonwealth University Faculty Grants‐in‐Aid.]

Frequency discrimination in regions of normal and impaired sensitivity.

Psychometric functions for frequency discrimination were obtained from both normal-hearing and hearing-impaired listeners at 300, 1200, and 3000 Hz. Difference limens for frequency (DLFs) were derived from those psychometric functions after repeated practice sessions, ensuring that all listener's results represented levels of optimum performance. Practice effects were found to be considerable in many listeners. The psychometric functions for all listeners, both normal-hearing and hearing-impaired, were well described by linear functions of performance level (d') and frequency difference (Hz), which passed through the zero origin. Those listeners with normal sensitivity thresholds at all frequencies exhibited the most acute frequency discriminations. Listeners with sensitivity losses at high frequencies exhibited larger DLFs, even at lower test frequencies where sensitivity thresholds were normal.

Frequency discrimination at 1200 Hz in the presence of high-frequency masking noise.

To evaluate the possible contribution of high-frequency hearing to frequency discrimination at lower frequencies, pure-tone frequency discrimination tasks at 1200 Hz, with and without high-frequency masking noise, were performed by two highly practiced normal-hearing listeners. Signal levels ranged from 10 to 80 dB SPL. Spectra of the three high-frequency masking noises ranged from 1.8 to 2.1 kHz, from 2.8 to 3.1 kHz, and from 4 to 8 kHz. The frequency difference limen (DLF) intensity functions, describing DLFs as a function of signal level which were obtained in the presence of high-frequency masking noise, were essentially the same as the DLF intensity functions obtained without masking noise. These results indicate that high-frequency hearing is not necessary for acute frequency discriminations at low and at middle frequencies.

The relation between sensation level and frequency discrimination under conditions of informational masking

Previous studies of resolution of the frequency, intensity, and duration of individual elements of word-length tonal sequences have shown major degrading effects of trial-to-trial stimulus uncertainty. The same degree of stimulus uncertainty was recently found to cause elevations of detection thresholds of 40–50 dB above the levels expected from temporal masking [Watson, Kelly, and Mellen, J. Acoust. Soc. Am. 66, S83 (1979)]. We referred to the effect of stimulus uncertainty on the threshold of individual elements of patterns as informational masking. In the current experiment, we investigated the relation between the amount of informational masking and frequency discrimination for the same target tones. Detection and discrimination trials were interleaved in an adaptive psycho-physical sequence to measure observers' detection thresholds and frequency difference limens for a 45-ms, 1688-Hz target tone, presented 90 ms after the onset of a 10-tone pattern. Context (non-target) tones varied between 25–45, 35–55, or 65–85 dB SPL in level, 300–3000 Hz in frequency, and 25–105 ms in duration. Target-tone levels on frequency discrimination trials were 35, 45, 55, or 75 dB. Informational masking was found to increase 10 dB for each 10-dB increase in context level, essentially as does energy masking, although the source of the two effects is clearly different. At all negative sensation levels, the threshold values of ΔF were those required to shift target-tone frequency beyond the limits of the pattern bandwidth. Frequency discrimination improved systematically as current “effective” sensation level increased from 0.0 to 40–50 dB, yielding a range of just-detectable values of Δf from over 1000 to about 20 Hz. The relation between sensation level and frequency discrimination is thus quite similar for informational and energy masking. [Work supported by NIH.]

Do the Blind Hear Better? Investigations on Auditory Processing in Congenital or Early Acquired Blindness I. Peripheral Functions

18 normally hearing blind students and 18 matched normal-sighted students were compared with regard to differential auditory sensitivity and acoustic reflex thresholds. Evaluation of interaural time differences for directional hearing and acoustic reflex thresholds showed no differences between the groups. The difference limen for intensity when interpreted by the SISI test, and the difference limen for frequency revealed a slightly optimizing trend in the blind group. However, differences missed statistical significance as to difference limen for intensity and, in the 3-orange, also as to difference limen for frequency. Any superior hearing of the blind was not mirrored by a corresponding superiority in peripheral supraliminal performance.

Frequency discrimination of the surface-feeding fishAplocheilus lineatus ? A prerequisite for prey localization?

1. The ability of the topminnowAplocheilus lineatus to discriminate water surface waves of different frequencies within the range 10 Hz to 150 Hz — which is relevant to prey identification and location — has been studied. 2. Frequency discrimination depends on reference frequency (RF). The frequency difference limen DL at a significance level of 5% was 2 Hz at 20 Hz, 3.7 Hz at 40 Hz, 6.5 Hz at 80 Hz and 17.2 Hz at 120 Hz RF (Fig. 2). The average Weber fraction DL/f was 0.03.

Amplitude modulation thresholds for the parakeet (Melopsittacus undulatus)

Parakeets were tested for the ability to detect sinusoidal amplitude modulation of broad band noise. Instrumental avoidance conditioning and a psychophysical modified method of limits procedure were used to measure the threshold for detecting amplitude modulation at 10 modulation frequencies between 2 and 2,048 Hz. Below about 40 Hz, modulation threshold is independent of modulation rate and noise level. Above 40 Hz, modulation threshold decreases with modulation frequency at the rate of 3 dB/ octave. These results are somewhat different from amplitude modulation functions in humans suggesting different degrees of temporal resolving power in birds and humans. Thresholds for changes in modulation rate are 1–2 orders of magnitude higher than pure tone frequency difference limens.

Theory of frequency modulation detection for low modulation frequencies.

There is general agreement that the frequency difference limen measured in a two-sine-tone frequency discrimination experiment is smaller than that measured in a frequency modulation (FM) experiment. We present a model of frequency modulation detection for low modulation frequencies, within the framework of signal detection theory, which accounts well for the observed difference between frequency discrimination experiments and FM detection experiments. The FM detection model also predicts psychometric functions for detection of FM with different modulation waveforms. FM detection experiments with square, sine, trapezoid, and triangle FM are in reasonable agreement with the model predictions.

Discrimination of brief tones by humans with temporal lobe damage

A number of recent Russian investigations [A. V. Baru and T. A. Karaseva, The Brain and Hearing, Consultants Bureau, New York, (1972)] reported that dogs and humans with large auditory cortex lesions, exhibit significant increases in both their absolute detection thresholds as well as their frequency difference limens for tonal signals which are shorter than 16 ms in duration. In contrast, more recent studies [J. L. Cranford, J. Acoust. Soc. Am. 65, 1573–1575 (1979)] with cats suggest that the cortex, rather than being essential for detecting the simple presence of brief tones, may be more important for discriminating qualitative differences in such sounds. While having normal absolute detection thresholds for brief tone pulses, cats with auditory cortex lesions, in comparison to unoperated controls, do exhibit elevated frequency difference limens. In order to investigate the inter‐species generality of these new animal findings, we recently presented the same series of brief tone tests to two human patients with temporal lobe lesions (one left‐sided CVA, and one right‐sided glioblastoma case). The results with the two patients are remarkably similar to those obtained with cats. Both patients exhibited normal absolute detection thresholds for brief tones in combination with significantly elevated frequency difference limens. [Work supported by Deafness Research Foundation and NINCDS.]

Difference limens for fundamental frequency contours in sentence level stimuli

Two experiments were performed to measure the difference limen (DL) for fundamental frequency using sentence level speech stimuli. In experiment I a male reader recorded five sentences. They were analyzed and synthesized using a linear prediction vocoding system. Part of each sentence was varied in 5-Hz steps above and below the original contour. Sentences (original contour and one of the changed contours) were presented to listeners using the method of constant stimuli. DL's ranged from less than 5 to 15 Hz depending on the sentence. There was no significant difference between DL's as a function of direction of change (up or down from original). Experiment II was an attempt to determine if the variability in DL due to sentences was a function of differences in placement of the frequency change. The changes in fundamental frequency began and ended at voiceless portions in all sentences. The median DL for the speaker also used in experiment I was 12 Hz. No significant differences as a function of sentences were found. For another speaker the DL's were over 15 Hz.

Pure‐tone and FM paradigms in the measurement of frequency discrimination

Previous research has shown that the choice of test method can dramatically effect obtained frequency difference limens (Δf). We measured Δf at several frequencies from 125 to 6000 Hz, using both 300‐ms pure‐tone bursts and tones which were frequency‐modulated at 8 Hz. Stimulus levels were approximately 40 dB SL, and 58 experimentally naive subjects were tested with a 4IFC adaptive procedure. The results confirm and extend some of the findings of Moore [Br. J. Audiol. 10, 17 (1976)]. Compared to the pure‐tone results, the FM tests revealed less variability, fewer learning effects, and lower Δf's. Discriminability measures across the two conditions were significantly correlated only at and above 2 kHz. Both paradigms revealed significant individual differences and subject‐by‐frequency interactions. For either paradigm, log Δf can be well described as a linear or a power function of log f, or as a linear function of f½. The results suggest (contrary to some earlier findings) that the FM paradigm may be a superior way to measure the frequency DL, particularly in clinical settings. [Research supported by NIH.]

External Electrical Stimulation of the Cochlea: Clinical, Psychophysical, Speech-Perceptual and Histological Findings

Our progress towards the development of a particular form of cochlear implant for the totally deaf is described. A single channel stimulation at the round window or promontory is used. This involves a minimum of surgical intervention and infective risk, preserves the possibility of remission and allows the application of later developments. The signal used for stimulation is designed to be matched both to the deaf lip-reader's needs and to his new, restricted, auditory ability. This is done by concentrating on the acoustic pattern components of speech which carry intonation and voiced-voiceless information. Surgical electrophysical, psychoacoustic and speech perceptual aspects of our work with twelve patients are described. The tests involve responses, for example, relating to: threshold for sinusoids; frequency difference limens; periodic -aperiodic discrimination; stress placement; and consonant labelling using combined visual and electrical inputs. Relatively extensive measurements were made with six patients. Significant individual differences were found and the sets of responses provide an essential basis for an appraisal of the potential usefulness of our work to the individual patient. Possible reasons for the individual differences are discussed. A brief indication is given of the techniques which we have developed for the future speech training and speech production evaluation of patients with electro-cochlear voice monitoring. The final section of our paper mentions our histological investigation of the effects of this type of stimulation in the guinea pig.

Behavior and psychophysics of hearing in birds

Psychophysical investigations of hearing in a number of avian species over the last decade have added significantly to our knowledge of hearing capabilities characteristic of this vertebrate group. Behavioral measures of absolute auditory sensitivity in a wide variety of bird species show a region of maximum sensitivity between 1 and 5 kHz with a rapid decrease in sensitivity at higher frequencies. On the basis of this general measure, birds fall between two other major vertebrate groups: reptiles and mammals. Discrimination and masking data from birds include measures of frequency, intensity, and duration difference limens; critical ratios, critical bands, and psychophysical tuning curves. There are also data on temporal summation, temporal resolving power, and temporary threshold shift from noise exposure. Taken together these data suggest that, in the region of 1–5 kHz, birds show a level of hearing sensitivity similar in most respects to that found for the most sensitive members of the class Mammalia with avian performance clearly inferior above and below this range of frequencies. Possible exceptions to this general picture must include the echolocating oilbird and growing evidence that pigeons are sensitive to infrasound at moderate intensity levels. The relation among critical ratio, critical band, and intensity difference limen measures in the parakeet is similar to that described for the human, but the pattern of masking as a function of frequency is dramatically different from that observed in mammals. Examples of a correspondence between hearing sensitivity and vocalizations can be demonstrated in a number of species.

Phase-locking in goldfish saccular nerve fibres accounts for frequency discrimination capacities.

THE auditory systems of fishes have been of particular interest in the development of auditory theory because a mechanical analysis of frequency had seemed unlikely on the basis of the ear's rather undifferentiated structural organisation1. Thus, at the level of the 8th nerve, information about stimulus frequency is thought to be coded in the temporal structure of neural activity rather than in across-fibre distributions of impulse rate2–4. However, recent neurophysiological data show that a limited peripheral frequency analysis occurs in several species5–7 and some workers have suggested that this may form the basis for psychophysical demonstrations of filtering8–10 and for frequency discrimination in the fish auditory system11,12. Such a ‘place principle’ of frequency analysis and pitch perception is presently generally believed to originate in the phyletic series among the amphibia and to reach the greatest degree of elaboration with the mammals13. We report here an experiment which helps to resolve this question of frequency codings by evaluating the adequacy of the temporal hypothesis to account for the behaviourally measured frequency difference limens for the goldfish11,12. A simple form of the temporal coding hypothesis states that the reciprocal of stimulus frequency (period) is represented in 8th nerve fibres as the distribution of time intervals between phase-locked impulses4. A discrimination between two different period lengths may thus entail a decision as to whether the two resulting distributions of neural interspike intervals are the same or different. Clearly, the performance on such a discrimination would be limited by the variability of the distributions, or the accuracy with which neural impulses are phase-locked to the stimulating waveform. We demonstrate here that the neurones which phase-lock best at any given frequency show degrees of phase-locking accuracy that would be expected if the central nervous system were processing the input optimally.

Imitation of synthesized vowels by preschool children

Preschool children (mean age of 4 years, 6 months) imitated 15 synthesized formant patterns representing both English and alien vowels. The first three formant frequencies of the imitations were measured from spectrograms. Generally, the children were successful in reproducing the relative formant patterns of the stimuli, although they tended to produce alien vowels as proximal English vowels. The reliability of imitation was better for stimuli modeled on English vowels. Combined with earlier results for men, women, and six-year-old children, the data are consistent with nearly straight isovowel lines in the F– F and F–F planes. The most linear relationship, with a good separation of the mean data for four age–sex groups, was noted for vowel [i]. The results for preschoolers, like those for adult subjects, indicate that the precision of vowel imitation is limited as much, if not more, by auditory resolution as by motor-articulatory performance, because the formant variability in imitation is commensurate with data on difference limens for vowel formant frequency.

A developmental study of speech production: Data on vowel imitation and sentence recitation

Ten subjects in each of five groups (4-, 6-, 10-, and 12-year-old children as well as young adults) participated in a study involving the imitation of 15 synthesized vowels and the repetition of three simple sentences. This report emphasizes spectrographic data for the two youngest groups and the adult group. In the vowel imitation task, the subjects could recreate the relative formant structure of the targets, but the children in particular tended to produce allen (non-English) vowels as the proximal English vowels. Plotting of the formant data for different age-sex groups in the F1−F2 plane revealed linear orientations, i.e., isovowel lines for unnormalized data. Standard deviations for the first two formant frequencies of the imitations were commensurate with published data on difference limens for formant frequency. The sentence repetitions were used primarily to evaluate the stability of temporal patterns. Both the inter- and intrasubject standard deviations of temporal segments tended to be larger for the children than for the adults, but even some 4-year-olds occasionally fell within the adult range of variability. [Work supported by NINCDS.]

Effect of hemispherectomy on frequency difference limens for the two ears

Hemispherectomized subjects display a strong ear dominance for the pitch mixture of dichotic chords: The tone presented to the ear ipsilateral to the removed hemisphere is less salient in the dichotic sound image than is the contralateral ear tone [R. Efron, M. Dennis, and J. E. Bogen, J. Acoust. Soc. Am. 60, Suppl. No. 1, 550 (1976)]. Two possible causes for similar ear dominance seen in normal subjects are an asymmetry in the transducer properties of the two ears or an asymmetry in the way two basically symmetrical inputs are combined centrally [E.W. Yund and R. Efron, J. Acoust. Soc. 62, 607–617 (1977)]. Evidence for the former derives from the finding that the dominant ear tends to have a superior frequency resolving capacity [P.L. Divenyi, R. Efron, and E.W. Yund, J. Acoust. Soc. Am. 62 624–632 (1977)]. Evidence for the latter derives from the present experiments which show that, in hemispherectomized subjects, there is no difference between the frequency dl in the two ears. The results suggest the existence of an efferent pathway from the cortex which can influence the way pitch information from the two ears is combined (weighted). [Work supported by the Veterans Administration.]