Do we parse the background into separate streams in the cocktail party?

Abstract

In the cocktail party situation, people with normal hearing usually follow a single speaker among multiple concurrent ones. However, there is no agreement in the literature as to whether the background is segregated into multiple streams/speakers. The current study varied the number of concurrent speech streams and investigated target detection and memory for the contents of a target stream as well as the processing of distractors. A male-spoken target stream was either presented alone (single-speech), together with one male-spoken (one-distractor), or a male- and a female-spoken distractor (two-distractor). Behavioral measures of target detection and content tracking performance as well as target- and distractor detection related ERPs were assessed. We found that the detection sensitivity and the target N2b amplitude decreased whereas the P3b amplitude increased from the single-speech to the concurrent speech streams conditions. Importantly, the behavioral distractor effect differed between the conditions with one- vs. two-distractor (distraction by the female speaker was lower than that of the male speaker in either condition) and the target N2b elicited in the presence of two distractors was significantly smaller than that elicited in the presence of one distractor. Further, the voltage in the N2b time window significantly differed between the one- and two-distractor conditions for the same (M2) speaker. These results show that speech processing was different in the presence of one vs. two distractors, and thus, the current data suggest that the two background speech streams were segregated from each other.

Do we parse the background into separate streams in the cocktail party?

“…Event-related potentials measured at the Pz electrode are shown on Figure 2. The scalp distributions of the target N2 and P3 show maximal amplitude for both components over parietal scalp locations (Figure 3), as was also seen in our previous studies (Szalárdy et al, 2018(Szalárdy et al, , 2019(Szalárdy et al, , 2020a.…”

“…In a representative example, the mean difference was calculated between target and distractor events: the mean difference was 2.348 s (SD: 1.722 s, min: 0.013 s, max: 7.535 s). Distractor articles (but targets not) also contained 19-26 syntactic violations (M = 20.5, SD = 1.4), which served for control purposes, as Szalárdy et al, 2018Szalárdy et al, , 2020a found that when participants follow one of two concurrent speech streams, syntactic violations within the unattended stream do not elicit the syntax-violation related ERP components. Therefore, syntactic violations could be used to indicate whether the non-target stream(s) were attended or not.…”

“…EEG recording and analysis were identical to Szalárdy et al (2018Szalárdy et al ( , 2020a. Continuous EEG was recorded (1 kHz sampling rate and 100 Hz online low-pass filter) from a few seconds before the beginning to a few seconds after the end of the stimulus blocks using a BrainAmp DC 64-channel EEG system with actiCAP active electrodes (Brain Products GmbH, Gilching, Germany).…”

“…We hypothesized that performance (both target detection and recognition performance) will be lower in the conditions with distractor speech streams compared to the single-speech condition. Concurrently, based on previous studies showing that the amplitude of the N2/P3 amplitudes to target events increase with increasing task demand (Isreal et al, 1980;Polich, 2007;Szalárdy et al, 2020a), we hypothesized that the N2/P3 elicited by target numerals will be larger in the presence of distractor speech compared to the single-speech condition. By using two distractor speech streams delivered by speakers of different gender, we aimed to provide acoustically sufficiently distinct stimuli to promote segregation of the two background streams.…”