Article Figures & Data
Figures
- Figure 1.
Stimulus paradigm and response from the EEG-TFS sensitivity measurement. A, The stimulus is a 1.5-s-long, 500-Hz pure tone that is amplitude modulated at 40.8 Hz. The red color represents the sound in the right ear, whereas the blue stands for the sound in the left ear. In the figure, the stimulus in the right ear leads in time till 0.98 s (indicated by the red segment of zoomed-out view of the stimulus), after which the ITD shifts in polarity, i.e., the stimulus in the left ear takes the lead. The ITD jump occurs when the stimulus amplitude is zero to minimize the involvement of monaural cues (pointed out by the dashed arrow). B, Averaged evoked response potential (ERP) from all trials across 42 subjects in “ITD = 540 μs” condition from Cz electrode. The red dashed line indicates where the ITD switched polarity, which resulted in N1 and P2 responses (denoted by red dots). C, ITC spectrogram of the EEG response, averaged across 42 subjects, with the colormap indicating the ITC. Robust ASSRs can be seen around the AM frequency of 40.8 Hz. There are also salient responses time locked to the stimulus onset, offset, and importantly, to the ITD jump. D, The average time course of the ITC for frequencies below 20 Hz is shown for each ITD jump condition. The response evoked by the shift in the ITD polarity increases monotonically with the size of the ITD jump, confirming that the response is parametrically modulated by TFS-based processing.
- Figure 2.
Measured versus predicted thresholds based on lapse rate. A, Measured versus predicted ITD detection thresholds. B, Measured versus predicted FM detection thresholds. The significant contribution of nonsensory factors is apparent, especially for the poorer performers.
- Figure 3.
Measured versus predicted AM thresholds based on lapse rate. The thresholds are the average detection thresholds of AM tones at 4 and 8 kHz. The significant contribution of nonsensory factors is apparent.
- Figure 4.
Individual EEG ITC (averaged under 20 Hz) values as a function of the jump size of the ITD. The ITC increases with the ITD for almost all subjects. Robust responses above noise floor are detected for most subjects for the “ITD = 180 μs” condition. Interestingly, individual differences present at 180 μs persist even at 540 μs despite the ITD jump being obviously perceptible and the response amplitude appearing to saturate.
- Figure 5.
FFR to the probe-forward-masker-probe stimulus sequence for an individual subject. The top row (green trace) represents the differential response across two stimulus polarities, whereas the bottom row (blue trace) represents the summed response across two stimulus polarities. The first boxed segments in both rows (red, dashed box, labeled d1 or s1) reflect the raw response to the probe tone, which is likely a mixture of neural and preneural responses (e.g., CM), whereas the second boxed segments in both rows (red, dashed box, labeled d2 or s2) is the adapted response after forward masking. For d2 and s2, the preneural (e.g., CM) component is expected to be intact, whereas the neural response is attenuated by forward masking (because of a very short 1-ms gap). The forward masker only partially suppresses the responses, suggesting a strong preneural contribution to d1 and s1. The weaker residuals obtained by subtraction, i.e., (d1 – d2) and (s1 – s2) are likely purely neural.
- Figure 6.
Frequency-domain representations of the d1 (A), d1–d2 (B), s1 (C), and s1–s2 (D) segments from Figure 5, but averaged across subjects. Forward masking partially attenuates both the 500-Hz component of d1 response, and the 1000-Hz component of the s1 response, suggesting that both responses reflect a mix of preneural and neural sources.
- Figure 7.
Model prediction of the ITD detection thresholds, based on the combination of lapse rate and slope (60- to 180-μs condition; A), or the combination of lapse rate and EEG latency (B). Please refer to Table 1 for the variance explained by each factor.
- Figure 8.
A sample of published reports of FM detection thresholds for comparison (Shower and Biddulph, 1931; Harris, 1952; Moore and Sek, 1996; He et al., 1998; Buss et al., 2004; Strelcyk and Dau, 2009; Ruggles et al., 2011; Grose and Mamo, 2012; Whiteford and Oxenham, 2015; Whiteford et al., 2017; Parthasarathy et al., 2020; Lelo de Larrea-Mancera et al., 2020). Error bar is 1 SD. The size of the dot represents the number of subjects (Whiteford and Oxenham (2015) has the most subjects; N = 100). Stimulus parameters such as stimulus level, carrier frequency, and modulation frequency in the cited studies are similar to those used in the current study, with slight differences (Strelcyk and Dau, 2009; Ruggles et al., 2011, used carrier at 750 Hz). Some threshold values are approximate from figures [e.g., mean and SD had to be estimated based on median and range in the box whisker plots in Whiteford and Oxenham (2015) and Whiteford et al. (2017)]. The mean and SD from the young and middle-aged group from Grose and Mamo (2012) were combined to generate a single data point. Some authors expressed the threshold in terms of ΔF/Fc, where ΔF is frequency deviation, and Fc is the carrier frequency. Moore and Sek (1996) used ΔF that was in two directions, i.e., peak-peak. Subjects from some studies were highly experienced in psychoacoustic tasks hence the thresholds were very low/good. Whiteford and Oxenham (2015) and Whiteford et al. (2017) obtained thresholds that fall in the lower end of the results of the current study from a very large number of subjects. This may be because their subjects were younger NH listeners and the stimuli were presented diotically and dichotically instead of monaurally.
- Figure 9.
A sample of published reports of ITD detection thresholds for comparison (Klumpp and Eady, 1956; Zwicker, 1956; Hershkowitz and Durlach, 1969; Henning, 1983; Dye, 1990; Bernstein and Trahiotis, 2002; Strelcyk and Dau, 2009; Hopkins and Moore, 2010; Grose and Mamo, 2010; Brughera et al., 2013). Error bar is 1 SD. The size of the dot represents the number of subjects (the current study has the most subjects; N = 36). Stimulus parameters such as level and carrier frequency in the cited studies are similar to those used in the current study, with slight differences [Strelcyk and Dau (2009) used carrier at 750 Hz]. Note that some threshold values were extracted approximately from figures rather than direct numerical reports. Some of the studies used stimuli with the leading ear switching from one side to the other (labeled “dynamic,” marked in green color), whereas others presented an ITD only in the target intervals, with the reference being the midline (labeled “static,” marked in blue color). Note that the values from Hershkowitz and Durlach (1969) and Brughera et al. (2013) were halved since the authors used ITD/2 in each interval. The mean and SD from young and middle-aged cohort from Grose and Mamo (2010) were combined to generate a single data point. Subjects from some studies were highly experienced in psychoacoustic tasks.
Tables
- Table 1
Model prediction of the behavioral ITD detection thresholds, with factors including the nonsensory score, EEG latency, and EEG slope
Predictor Variance explained Nonsensory score 37.48% EEG latency 10.03% EEG slope 9.28% Explained 56.79% Unexplained 43.21% The variations accounted for by the nonsensory score are more than three times as by either one of the two EEG metrics. Together, more than half the variance can be explained.
In this issue
