Spiking network optimized for word recognition in noise predicts auditory system hierarchy

Models optimized for real-world tasks reveal the necessity of precise temporal coding in hearing

Neurons encode information in the timing of their spikes in addition to their firing rates. Spike timing is particularly precise in the auditory nerve, where action potentials phase lock to sound with sub-millisecond precision, but its behavioral relevance is uncertain. To investigate the role of this temporal coding, we optimized machine learning models to perform real-world hearing tasks with simulated cochlear input. We asked how precise auditory nerve spike timing needed to be to reproduce human behavior. Models with high-fidelity phase locking exhibited more human-like sound localization and speech perception than models without, consistent with an essential role in human hearing. Degrading phase locking produced task-dependent effects, revealing how the use of fine-grained temporal information reflects both ecological task demands and neural implementation constraints. The results link neural coding to perception and clarify conditions in which prostheses that fail to restore high-fidelity temporal coding could in principle restore near-normal hearing.

Many but not all deep neural network audio models capture brain responses and exhibit correspondence between model stages and brain regions

Models that predict brain responses to stimuli provide one measure of understanding of a sensory system and have many potential applications in science and engineering. Deep artificial neural networks have emerged as the leading such predictive models of the visual system but are less explored in audition. Prior work provided examples of audio-trained neural networks that produced good predictions of auditory cortical fMRI responses and exhibited correspondence between model stages and brain regions, but left it unclear whether these results generalize to other neural network models and, thus, how to further improve models in this domain. We evaluated model-brain correspondence for publicly available audio neural network models along with in-house models trained on 4 different tasks. Most tested models outpredicted standard spectromporal filter-bank models of auditory cortex and exhibited systematic model-brain correspondence: Middle stages best predicted primary auditory cortex, while deep stages best predicted non-primary cortex. However, some state-of-the-art models produced substantially worse brain predictions. Models trained to recognize speech in background noise produced better brain predictions than models trained to recognize speech in quiet, potentially because hearing in noise imposes constraints on biological auditory representations. The training task influenced the prediction quality for specific cortical tuning properties, with best overall predictions resulting from models trained on multiple tasks. The results generally support the promise of deep neural networks as models of audition, though they also indicate that current models do not explain auditory cortical responses in their entirety.

Auditory perception architecture with spiking neural network and implementation on FPGA

Spike-based perception brings up a new research idea in the field of neuromorphic engineering. A high-performance biologically inspired flexible spiking neural network (SNN) architecture provides a novel method for the exploration of perception mechanisms and the development of neuromorphic computing systems . In this article, we present a biological-inspired spike-based SNN perception digital system that can realize robust perception. The system employs a fully paralleled pipeline scheme to improve the performance and accelerate the processing of feature extraction. An auditory perception system prototype is realized on ten Intel Cyclone field-programmable gate arrays, which can reach the maximum frequency of 107.28 MHz and the maximum throughput of 5364 Mbps. Our design also achieves the power of 5. 148 W/system and energy efficiency of 845.85 μJ. Our auditory perception implementation is also proved to have superior robustness compared with other SNN systems. We use TIMIT digit speech in noise in accuracy testing. Result shows that it achieves up to 85.75% speech recognition accuracy under obvious noise conditions (signal-to-noise ratio of 20 dB) and maintain small accuracy attenuation with the decline of the signal-to-noise ratio. The overall performance of our proposed system outperforms the state-of-the-art perception system on SNN.

Encoding of speech in convolutional layers and the brain stem based on language experience

Comparing artificial neural networks with outputs of neuroimaging techniques has recently seen substantial advances in (computer) vision and text-based language models. Here, we propose a framework to compare biological and artificial neural computations of spoken language representations and propose several new challenges to this paradigm. The proposed technique is based on a similar principle that underlies electroencephalography (EEG): averaging of neural (artificial or biological) activity across neurons in the time domain, and allows to compare encoding of any acoustic property in the brain and in intermediate convolutional layers of an artificial neural network. Our approach allows a direct comparison of responses to a phonetic property in the brain and in deep neural networks that requires no linear transformations between the signals. We argue that the brain stem response (cABR) and the response in intermediate convolutional layers to the exact same stimulus are highly similar without applying any transformations, and we quantify this observation. The proposed technique not only reveals similarities, but also allows for analysis of the encoding of actual acoustic properties in the two signals: we compare peak latency (i) in cABR relative to the stimulus in the brain stem and in (ii) intermediate convolutional layers relative to the input/output in deep convolutional networks. We also examine and compare the effect of prior language exposure on the peak latency in cABR and in intermediate convolutional layers. Substantial similarities in peak latency encoding between the human brain and intermediate convolutional networks emerge based on results from eight trained networks (including a replication experiment). The proposed technique can be used to compare encoding between the human brain and intermediate convolutional layers for any acoustic property and for other neuroimaging techniques.

Quantitative models of auditory cortical processing

To generate insight from experimental data, it is critical to understand the inter-relationships between individual data points and place them in context within a structured framework. Quantitative modeling can provide the scaffolding for such an endeavor. Our main objective in this review is to provide a primer on the range of quantitative tools available to experimental auditory neuroscientists. Quantitative modeling is advantageous because it can provide a compact summary of observed data, make underlying assumptions explicit, and generate predictions for future experiments. Quantitative models may be developed to characterize or fit observed data, to test theories of how a task may be solved by neural circuits, to determine how observed biophysical details might contribute to measured activity patterns, or to predict how an experimental manipulation would affect neural activity. In complexity, quantitative models can range from those that are highly biophysically realistic and that include detailed simulations at the level of individual synapses, to those that use abstract and simplified neuron models to simulate entire networks. Here, we survey the landscape of recently developed models of auditory cortical processing, highlighting a small selection of models to demonstrate how they help generate insight into the mechanisms of auditory processing. We discuss examples ranging from models that use details of synaptic properties to explain the temporal pattern of cortical responses to those that use modern deep neural networks to gain insight into human fMRI data. We conclude by discussing a biologically realistic and interpretable model that our laboratory has developed to explore aspects of vocalization categorization in the auditory pathway.

Two stages of bandwidth scaling drives efficient neural coding of natural sounds

Theories of efficient coding propose that the auditory system is optimized for the statistical structure of natural sounds, yet the transformations underlying optimal acoustic representations are not well understood. Using a database of natural sounds including human speech and a physiologically-inspired auditory model, we explore the consequences of peripheral (cochlear) and mid-level (auditory midbrain) filter tuning transformations on the representation of natural sound spectra and modulation statistics. Whereas Fourier-based sound decompositions have constant time-frequency resolution at all frequencies, cochlear and auditory midbrain filters bandwidths increase proportional to the filter center frequency. This form of bandwidth scaling produces a systematic decrease in spectral resolution and increase in temporal resolution with increasing frequency. Here we demonstrate that cochlear bandwidth scaling produces a frequency-dependent gain that counteracts the tendency of natural sound power to decrease with frequency, resulting in a whitened output representation. Similarly, bandwidth scaling in mid-level auditory filters further enhances the representation of natural sounds by producing a whitened modulation power spectrum (MPS) with higher modulation entropy than both the cochlear outputs and the conventional Fourier MPS. These findings suggest that the tuning characteristics of the peripheral and mid-level auditory system together produce a whitened output representation in three dimensions (frequency, temporal and spectral modulation) that reduces redundancies and allows for a more efficient use of neural resources. This hierarchical multi-stage tuning strategy is thus likely optimized to extract available information and may underlies perceptual sensitivity to natural sounds.

Inferring the basis of binaural detection with a modified autoencoder

The binaural system utilizes interaural timing cues to improve the detection of auditory signals presented in noise. In humans, the binaural mechanisms underlying this phenomenon cannot be directly measured and hence remain contentious. As an alternative, we trained modified autoencoder networks to mimic human-like behavior in a binaural detection task. The autoencoder architecture emphasizes interpretability and, hence, we "opened it up" to see if it could infer latent mechanisms underlying binaural detection. We found that the optimal networks automatically developed artificial neurons with sensitivity to timing cues and with dynamics consistent with a cross-correlation mechanism. These computations were similar to neural dynamics reported in animal models. That these computations emerged to account for human hearing attests to their generality as a solution for binaural signal detection. This study examines the utility of explanatory-driven neural network models and how they may be used to infer mechanisms of audition.

Multiscale temporal integration organizes hierarchical computation in human auditory cortex

To derive meaning from sound, the brain must integrate information across many timescales. What computations underlie multiscale integration in human auditory cortex? Evidence suggests that auditory cortex analyses sound using both generic acoustic representations (for example, spectrotemporal modulation tuning) and category-specific computations, but the timescales over which these putatively distinct computations integrate remain unclear. To answer this question, we developed a general method to estimate sensory integration windows-the time window when stimuli alter the neural response-and applied our method to intracranial recordings from neurosurgical patients. We show that human auditory cortex integrates hierarchically across diverse timescales spanning from ~50 to 400 ms. Moreover, we find that neural populations with short and long integration windows exhibit distinct functional properties: short-integration electrodes (less than ~200 ms) show prominent spectrotemporal modulation selectivity, while long-integration electrodes (greater than ~200 ms) show prominent category selectivity. These findings reveal how multiscale integration organizes auditory computation in the human brain.

Understanding Adaptive, Multiscale Temporal Integration In Deep Speech Recognition Systems

Natural signals such as speech are hierarchically structured across many different timescales, spanning tens (e.g., phonemes) to hundreds (e.g., words) of milliseconds, each of which is highly variable and context-dependent. While deep neural networks (DNNs) excel at recognizing complex patterns from natural signals, relatively little is known about how DNNs flexibly integrate across multiple timescales. Here, we show how a recently developed method for studying temporal integration in biological neural systems - the temporal context invariance (TCI) paradigm - can be used to understand temporal integration in DNNs. The method is simple: we measure responses to a large number of stimulus segments presented in two different contexts and estimate the smallest segment duration needed to achieve a context invariant response. We applied our method to understand how the popular DeepSpeech2 model learns to integrate across time in speech. We find that nearly all of the model units, even in recurrent layers, have a compact integration window within which stimuli substantially alter the response and outside of which stimuli have little effect. We show that training causes these integration windows to shrink at early layers and expand at higher layers, creating a hierarchy of integration windows across the network. Moreover, by measuring integration windows for time-stretched/compressed speech, we reveal a transition point, midway through the trained network, where integration windows become yoked to the duration of stimulus structures (e.g., phonemes or words) rather than absolute time. Similar phenomena were observed in a purely recurrent and purely convolutional network although structure-yoked integration was more prominent in the recurrent network. These findings suggest that deep speech recognition systems use a common motif to encode the hierarchical structure of speech: integrating across short, time-yoked windows at early layers and long, structure-yoked windows at later layers. Our method provides a straightforward and general-purpose toolkit for understanding temporal integration in black-box machine learning models.

Reconstruction of a Fully Paralleled Auditory Spiking Neural Network and FPGA Implementation

This paper presents a field-programmable gate array (FPGA) implementation of an auditory system, which is biologically inspired and has the advantages of robustness and anti-noise ability. We propose an FPGA implementation of an eleven-channel hierarchical spiking neuron network (SNN) model, which has a sparsely connected architecture with low power consumption. According to the mechanism of the auditory pathway in human brain, spiking trains generated by the cochlea are analyzed in the hierarchical SNN, and the specific word can be identified by a Bayesian classifier. Modified leaky integrate-and-fire (LIF) model is used to realize the hierarchical SNN, which achieves both high efficiency and low hardware consumption. The hierarchical SNN implemented on FPGA enables the auditory system to be operated at high speed and can be interfaced and applied with external machines and sensors. A set of speech from different speakers mixed with noise are used as input to test the performance our system, and the experimental results show that the system can classify words in a biologically plausible way with the presence of noise. The method of our system is flexible and the system can be modified into desirable scale. These confirm that the proposed biologically plausible auditory system provides a better method for on-chip speech recognition. Compare to the state-of-the-art, our auditory system achieves a higher speed with a maximum frequency of 65.03 MHz and a lower energy consumption of 276.83 μJ for a single operation. It can be applied in the field of brain-computer interface and intelligent robots.