Channel,temporal, and composite uncertainty in the detection and recognition of auditory and visual signals

In a 2AFC detection task, channel and temporal uncertainty were separately shown to produce decrements in the detectability of auditory and visual signals. The observed effect of channel uncertainty suggests that Ss are capable of using an “attentional” or “performance-compensation” strategy. This finding is discussed in relation to a “multiple-observation” model for channel uncertainty. Recognition measures were taken where appropriate and gave some support to the hypothesized attentional strategy. Under composite uncertainty conditions both types of uncertainty were presented simultaneously; the results indicated that any interaction effect was small, relative to the separate effects of channel and temporal uncertainty.     

Interval of time uncertainty in visual detection

The effects of increasing the length of an interval of continuous temporal uncertainty (ITU) on the detectability of visualsignals were investigated in a ‘Yes/No’ detection situation. Ss were uncerlain about when a signal might occur within a given observalion interval, the duration of which was varied. Longer intervals of uncertainty resulted in a decrement in detectability, which was shown to be directly attributable to increased false alarm rates. It was suggested that observers have more opportunities for confusing signals with noise. The time course of detectability within a givenITU was also investigated, but there were no significant variations.     

A comparison of auditory monitoring performance in blind subjects with that of sighted subjects in light and dark

Two experiments were performed in which blind Ss, sighted Ss working in the dark, and sighted Ss working in the light were compared as to their efficiency on an auditory watchkeeping task. Absolute and differential auditory thresholds were measured in both experiments, and in the second experiment the groups also underwent a signal detection session under alerted conditions. There was some inconsistency as to relative performance of the sighted groups, but in both experiments the blind Ss were superior on the auditory watchkeeping task as to signals detected and effective sensitivity (d’). 1 This difference was not attributable to a difference in auditory sensitivity or to a criterion adopted for     

Frequency-selective detection at three signal amplitudes

Using the probe-signal method, frequency-response characteristics were obtained for four Os detecting signals nominally of only 1000Hz under three signal-amplitude conditions: E/N_o=10.0, 12.6, and 15.8. Detection of 1000-Hz signals decreased systematically with signal-amplitude decreases, but signals at 850 Hz and at 1150 Hz were detected at or near the chance level during all signal-amplitude conditions. From the detection performance data, the detectability was inferred for four probe-signal frequencies relative to that of 1000-Hz signals for each of the amplitude conditions. From the curvesofrelative detectability, hal[-detectability bandwidths were estimated. Neither the relative detectability curves nor the half-detectability bandwidths indicate any systematic change in the frequencyselectivebehavior of the auditory O across the signal amplitudes used.     

The role of auditory memory traces in attention to frequency

Three cued signal detection experiments demonstrated a role for auditory memory traces in frequency selectivity. The extent to which the cue predicted the signal frequency affected the size of the advantage for signals at the cue frequency over those at distant frequencies when the cue-signal gap was 10 sec but not when it was 1 sec. Detection of occasional signals presented at uncued frequencies was enhanced when they matched the frequency of cues from recent trials. With "relative" cues, which were usually followed by signals at the musical fifth above the cue frequency, performance on occasional signals at the cue frequency was enhanced relative to other unexpected frequencies. These results suggest that, regardless of the listener's expectations and intentions, the detectability of a signal is enhanced if its frequency matches an existing memory trace. One form of voluntary attention to frequency may involve maintaining traces that would otherwise slowly decay.     

Detection and recognition of auditory patterns

The detection/recognition theorem (Starr, Metz, Lusted, & Goodenough, 1975) provides a signal detection theory prediction of an observer's ability to identify one of a set of possible signals on the basis of the observer's ability to detect whether any signal is presented. The present findings show that the theorem can be applied to complex auditory stimuli in a fashion that is not merely a trivial extension of the results obtained with simple auditory stimuli.     

Nonlinguistic auditory capabilities in aphasia

Nonlinguistic auditory capabilities were assessed through psychophysical tests in 11 left-CVA aphasic, four right-CVA nonaphasic, and eight normal male subjects selected from the same age group. The tests included frequency discrimination, gap detection, gap discrimination, frequency sweep discrimination, assessment of the magnitude of the frequency uncertainty effect in the detectability of tones in noise, and assessment of frequency selectivity through simultaneous masked thresholds. Results of these tests were compared to measures of auditory comprehension obtained from the Boston Diagnostic Aphasia Examination, the Porch Index of Communicative Ability, and the Token Test. Nonlinguistic auditory performance of the three subject groups differed significantly from each other. For the left-CVA subjects, frequency sweep discrimination, frequency discrimination, and the frequency uncertainty effect in tone-in-noise detection were the best predictors of verbal auditory comprehension. The right-CVA subjects displayed marked deficits with regard to all pitch-related tests. The findings stress the importance of considering the presence of nonlinguistic auditory dysfunctions when evaluating linguistic auditory capabilities in aphasia.     

Some properties of the olfactory system of man

Preliminary investigations into the effects of a 12-h diurnal variation suggested there was a 10-fold decrease in odor sensitivity from morning to evening. For a selected group of odorants (methyl, t-butyl, and phenyl isonitrile), the difference between detection and recognition thresholds in general was found to be 10-fold; however, for some Ss the difference was less. When Ss were blindfolded, no changes in odor sensitivity were noted. Similarly, imposing an auditory stress on the test situation did not significantly alter odor sensitivity in either experienced or naive Ss,     

Simultaneous detection and recognition of chromatic flashes

Studies of simultaneous detection and recognition were performed to test alternative models of the detection process, signal detection theory and low-threshold theory. Sensitivity in a detection experiment was independent of whether the type of signal (red or green light flash) was known in advance, because only one type of s trial was possible, or was unknown because either stimulus could occur. When a recognition judgment was added to either a binary or rating-scale detection response, Ss were able to report the nature of the stimulus at better than chance levels even when they indicated that the stimulus was not detected. Since such performance occurred when Ss used detection responses likely to have been given only in the nondetect state, the data lead to the rejection of low-threshold theory.     

Cuing mechanisms in autitory signal detection

Detection of auditory signals under frequency uncertainty can be improved by presenting cues to the listeners. Since various cues have been found to differ in effectiveness, three conceivable mechanisms were considered which might account for these differences. Cuing might reduce the number and/or width of the employed auditory filters or listening bands. Also, cues could modulate the precision of frequency tuning of the filters. Psychometric functions were collected in a detection experiment with frequency uncertainty employing three kinds of cues: pure tones whose frequency was identical to that of the signal (iconic cues), complex tones with a missing fundamental equal to the signal (complex cues), and pure tones with a certain frequency relation to the signal (relative cues). Compared with a no-cue condition, all cue types improved detection performance. Fitting models to the data suggests that in the no-cue condition as well as the complex-cue condition, multiple bands were utilized, and that the iconic and relative cues induced single-band listening. There is no indication that accuracy of frequency tuning was responsible for cue-efficiency differences.     

The relation between physical stimulus properties and the effect of foreperiod duration on reaction time

Seven subjects were used in an experiment on the relation between signal modality and the effect of foreperiod duration (EP) on RT. With visual signals the usually reported systematic increase of RT as a function of FP duration (1, 5 and 15 s) was confirmed; with auditory signals no difference was found between FP's of 1 and 5 s while the effect at 15 s was equivalent to that found at 5 s with the visual signal. The results suggest that besides factors such as time uncertainty the FP effect is also largely dependent on the arousing quality of the signal.     

The non-monotonicity of the psychometric function in recognition memory

The technique of using the signal detectability theory parameter d’ as a measure of trace strength in recognition memory is extended in a decision-making model that predicts that a S’s psychometric function lp(Y/s) plotted against stimulus intensity / will have a peak for intermediate values of trace strength when the S has a bias towards a “Yes” response. An experiment is reported where Ss were required to recognize previously presented nonsense syllables under a number of pay-off conditions designed to give different groups of Ss different response biases. The predictions of the model were confirmed. It is suggested that a S’s discriminability and response bias parameters may not be as stable as previous studies, that pool large samples of data, have indicated, and the results also provide support for a two-stage model of recognition, where a preliminary estimate of the trace strength of an item is obtained before the location of the response criterion is decided.     

Thomas and Legge’s matching hypothesis for

Thomas and Legge have proposed amatching hypothesis for two-alternative detection and recognition tasks. Given symmetric payoff, S is supposed to match his unconditional response probabilities to the presentation probabilities. More generally, even when payoff is asymmetric, S is supposed to keep his unconditional response probabilities constant across different discriminability levels. Time series analysis is applied here to the response sequences from two auditory amplitude recognition tasks to test both hypotheses. Experiment I obtains a single sequence from each of 16 Ss. Experiment II obtains two sequences at different discriminability levels from each of 24 Ss. Departures from the matching hypothesis are extremely gross in both cases. In addition, Experiment II shows clear changes in the response probabilities across discriminability conditions, though these are not systematic in direction across Ss.     

Operating characteristics and a priori probability of the signal

Two experiments in auditory signal detection produced changes in the operating characteristic as the a priori probability of signal occurrence, p(SN), was varied. The signal was a sinusoid of 1,000 Hz presented for 250 msec against a continuous background of noise. In Experiment 1 three values of p(SN)—0.25, 0.50, and 0.75—were paired with each of three signal intensities. In Experiment 2 the signal intensity was fixed and p(SN) was assigned values of 0.10, 0.25, 0.50, 0.75, and 0.90. On normal-normal coordinates, operating characteristics were fitted to the points obtained from the 4-point rating scale used by the listeners. Such operating characteristics may be specified by two parameters: d_s, an index of detectability related to d’, and m, its slope. While ds was found to be independent ofp(SN), m was found to be a joint function of p(SN) and signal intensity. These results are discussed against the background of the theory of signal detectability.     

A model for detection and recognition with signal uncertainty

A model for signal detectability suggested by Luce is extended to situations in which the observer is uncertain of some important characteristic of the signal, such as frequency. By making a single assumption concerning the observer's covert response behavior, two solutions are obtained corresponding to qualitatively different behavior. Decrements in detectability and in recognition with uncertainty are shown to be particular functions of discriminability and detectability of the stimuli in other situations. Relevant experimental data are considered.This work was supported by grant NSF-G5544 from the National Science Foundation. The author is indebted to R. Duncan Luce for critical discussion of the content and helpful suggestions concerning the form of this paper. Useful comments have been made by David M. Green, Francis W. Irwin, Roger M. Shepard, and W. P. Tanner, Jr.     

Effects of temporal uncertainty on watchkeeping performance

Two experiments were conducted to assess the relative effects of signal density and regularity on watchkeeping performance. In Experiment I, three levels of density (6, 24, and 96 signals/hr.) were combined factorially with three levels of variability (coefficients of variation of 0.01, 0.l0, and 1.00), and I0 Ss were assigned at random to each of the nine conditions. In Experiment II, five leveIs of density (6, 12, 24, 48, and 96 signals/hr.) were combined with the same three levels of variability, and 13 Ss were assigned to each condition. Each S monitored a visual “blinking-lights” display for an hour under instructions to detect and report the occurrence of certain “critical signals,” i.e., arrests of alternation of the lights. Response times (RT’s) to correctly detected signals in both experiments decreased as a linear function of logarithmic increases in signal density. An uncertainty metric, the signal surprisal due to density, was derived, and the watchkeeper’s RT was expressed as an increasing linear function of this measure of temporal uncertainty. Interpretation of these and 0 ther data support a functional, psychophysical approach to the study of watchkeeping behavior.     

Temporal and channel uncertainty in detection: A multiple observation analysis

A detection model is generalized to apply to certain multiple observation tasks and data from two auditory experiments are presented which suggest its validity. An observer’s ability to discriminate trials on which tone signals were present during a randomly selected t of n observations, from trials on which no signals were present, was shown to be consistent with the model over a wide range of n and t values. Special emphasis is placed on the performance decrement produced by increasing n relative to t. This “temporal uncertainty” effect is then related to “channel uncertainty” in experiments designed to study attentional processes.     

The time course of attention in a simple auditory detection task

What is the time course of human attention in a simple auditory detection task? To investigate this question, we determined the detectability of a 20-msec, 1000-Hz tone presented at expected and unexpected times. Twelve listeners who expected the tone to occur at a specific time after a 300-msec narrowband noise rarely detected signals presented 150-375 msec before or 100-200 msec after that expected time. The shape of this temporal-attention window depended on the expected presentation time of the tone and the temporal markers available in the trials. Further, though expecting the signal to occur in silence, listeners often detected signals presented at unexpected times during the noise. Combined with previous data, these results further clarify the listening strategy humans use when trying to detect an expected sound: Humans seem to listen specifically for that sound, while ignoring the background in which it is presented, around the time when the sound is expected to occur.     

Attention and auditory evoked responses to low-detectability signals

Comparisons were made between cortical evoked responses obtained under two conditions: (1) while Ss were reading, and (2) while they were attempting to count auditory signals. The amplitudes of evoked responses to low-detectability auditory stimuli were found to be approximately doubled when the Ss were required to count the number of stimuli, as compared to amplitudes recorded when they were reading. The duration of the response was also markedly increased. These increases in response amplitude and duration are considerably greater than those observed in earlierexperiments, where high-levelsignalswere used. Inter-S variability of the waveform of the averageevoked response was observed to be much less when the Ss counted the stimuli. In another experiment the level of the auditory signalwas varied over a range of approximately +4 to -4 decibels relative to the listeners’ behavioral thresholds. The per cent of signalswhich they counted varied from near-zero to 100, over this range, and the evoked response concurrently showed a variation from “unmeasureable” to approximately 8 microvolts.     

Feature saliency and recognition memory for schematic faces

Memory for schematic faces was studied using a recognition memory paradigm. Ss indicated whether two faces, separated by delays of 6, 12, or 18 sec, were the same or different. Theoretical interest centered on the effects of component feature saliency on response bias and detectability. The main results were that there were fewer misses and more false alarms on the more salient features. Analysis showed that the main effect of feature saliency was on response bias, not detectability.