Previous studies have shown that acoustic parameters such as frequency separation and presentation rate can influence the perceptual formation of auditory stream segregation, with the underlying neural mechanisms commonly attributed to changes in frequency tuning overlap. The population segregation model explains this phenomenon based on statistical properties of spatial distributions of neural population responses . However, under synchronous stimulation or large frequency separation conditions, inconsistencies remain between neural indices and perceptual outcomes; that is, auditory stream segregation may fail to emerge even when spatial separation is reduced. This discrepancy suggests that spatial statistics alone are insufficient to account for the phenomenon, and that the temporal structure of neural population responses may also play a critical role. The temporal coherence model addresses this issue from the perspective of temporal correlations; however, it primarily relies on constructively defined statistical correlation structures, and lacks a direct mapping to frequency tuning mechanisms. Therefore, there may exist an intermediate variable based on the statistical structure of neural population responses that bridges frequency tuning overlap and temporal structure representations, jointly influencing the formation of auditory stream segregation. To investigate this, we constructed a neural population response model based on the ABAB auditory stimulation paradigm. The model consists of two neuronal populations with distinct best frequencies, and neural responses are generated under a Poisson spiking assumption. Within this framework, we systematically analyzed how frequency tuning overlap influences the anti-phase alternating temporal structure of population responses. Furthermore, we introduced a low-dimensional proxy variable derived from population response differences to characterize the identifiability of this alternating structure, and further quantified its stability via temporal fluctuations. Under both idealized conditions and stochastic spiking regimes, our results show that a reduction in frequency tuning overlap simultaneously enhances both the identifiability and stability of this temporal structure, exhibiting a coupled relationship between the two. These findings suggest that frequency tuning overlap may influence auditory stream segregation by jointly modulating the identifiability and stability of neural population response temporal structure.
Wei Liu (Sun,) studied this question.