Article

Sequence Sensitivity of Neurons in Cat Primary Auditory Cortex

Coleman Laboratory, W.M. Keck Center for Integrative Neuroscience, University of California at San Francisco, San Francisco, CA 94143-0732, USA.
Cerebral Cortex (Impact Factor: 8.67). 01/2001; 10(12):1155-67. DOI: 10.1093/cercor/10.12.1155
Source: PubMed

ABSTRACT

Properties of sequence-sensitive neurons in primary auditory cortex of cats were explored in detail. Stimuli were sequences of two tones, in which the frequency and intensity of the first tone and the temporal separation between the first and second, or probe, tone were parametrically varied. After presentation of the first tone, the responses of 32 single units and 48 multiunits to the probe tone were found to be enhanced up to 140-5270% (median 340%) above the response obtained in the single-tone condition. Probe tone enhancement was induced from a considerable number of sequence conditions and depended on the frequency and intensity of the first tone and on the temporal separation between the onsets of the first and the probe tone. On average, the maximally enhanced response occurred when the first tone was 1 octave below or above the probe tone and its intensity was 14 dB louder than the probe tone. The most effective temporal separation of the tones for an enhancement effect was approximately 100 ms. The range of enhancing tones was largely outside the excitatory tuning curve of a neuron. Results extend previous findings of properties of sequence-sensitive neurons in the auditory cortex of echolocating bats and non-echolocating mammals, and suggest that sequence-sensitive neurons are quite common and involved in the cortical representation of spectrotemporal patterns of acoustic signals.

Download full-text

Full-text

Available from: Christoph E Schreiner
  • Source
    • "Neuroimaging studies have demonstrated that multiple cortical regions are involved in temporal sequence processing (Clegg et al., 1998; Mauk and Buonomano, 2004). Recent neurophysiology studies have shown that even neurons in primary visual cortex can learn to recognize and predict spatiotemporal sequences (Xu et al., 2012; Gavornik and Bear, 2014) and that neurons in primary visual and auditory cortex exhibit sequence sensitivity (Brosch and Schreiner, 2000; Nikolić et al., 2009). These studies suggest that sequence learning is an important problem that is solved by many cortical regions. "
    [Show abstract] [Hide abstract]
    ABSTRACT: The ability to recognize and predict temporal sequences of sensory inputs is vital for survival in natural environments. Based on many known properties of cortical neurons, a recent study proposed hierarchical temporal memory (HTM) sequence memory as a theoretical framework for sequence learning in the cortex. In this paper, we analyze properties of HTM sequence memory and apply it to various sequence learning and prediction problems. We show the model is able to continuously learn a large number of variable-order temporal sequences using an unsupervised Hebbian-like learning rule. The sparse temporal codes formed by the model can robustly handle branching temporal sequences by maintaining multiple predictions until there is sufficient disambiguating evidence. We compare the HTM sequence memory and other sequence learning algorithms, including the autoregressive integrated moving average (ARIMA) model and long short-term memory (LSTM), on sequence prediction problems with both artificial and real-world data. The HTM model not only achieves comparable or better accuracy than state-of-the-art algorithms, but also exhibits a set of properties that is critical for sequence learning. These properties include continuous online learning, the ability to handle multiple predictions and branching sequences, robustness to sensor noise and fault tolerance, and good performance without task-specific hyper-parameters tuning. Therefore the HTM sequence memory not only advances our understanding of how the brain may solve the sequence learning problem, but is also applicable to a wide range of real-world problems such as discrete and continuous sequence prediction, anomaly detection, and sequence classification.
    Full-text · Article · Dec 2015
  • Source
    • "Bartlett & Wang, 2005; Brosch & Scheich, 2008; Sadagopan & Wang, 2009). With both fast-and slowly-decaying synaptic adaptation , 16% of columns in the model showed facilitation – a lower proportion than the 40–60% values found in single-unit measurements in primates (Brosch & Schreiner, 2000; Bartlett & Wang, 2005; Brosch & Scheich, 2008). Irrespective of the time constant of synaptic adaptation, the magnitude of facilitation peaked in the lower range (110–140%), resembling the results from primate studies . "
    [Show abstract] [Hide abstract]
    ABSTRACT: Incoming sounds are represented in the context of preceding events, and this requires a memory mechanism that integrates information over time. Here, it was demonstrated that response adaptation, the suppression of neural responses due to stimulus repetition, might reflect a computational solution that auditory cortex uses for temporal integration. Adaptation is observed in single-unit measurements as two-tone forward masking effects and as stimulus-specific adaptation (SSA). In non-invasive observations, the amplitude of the auditory N1m response adapts strongly with stimulus repetition, and it is followed by response recovery (the so-called mismatch response) to rare deviant events. The current computational simulations described the serial core-belt-parabelt structure of auditory cortex, and included synaptic adaptation, the short-term, activity-dependent depression of excitatory corticocortical connections. It was found that synaptic adaptation is sufficient for columns to respond selectively to tone pairs and complex tone sequences. These responses were defined as combination sensitive, thus reflecting temporal integration, when a strong response to a stimulus sequence was coupled with weaker responses both to the time-reversed sequence and to the isolated sequence elements. The temporal complexity of the stimulus seemed to be reflected in the proportion of combination-sensitive columns across the different regions of the model. Our results suggest that while synaptic adaptation produces facilitation and suppression effects, including SSA and the modulation of the N1m response, its functional significance may actually be in its contribution to temporal integration. This integration seems to benefit from the serial structure of auditory cortex. © 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd.
    Full-text · Article · Mar 2015 · European Journal of Neuroscience
  • Source
    • "Brosch et al. (1999) observed maximal enhancement of multi-unit responses in A1 of anesthetized macaque monkeys when a pair of tones were delivered with a stimulus-onset-asynchrony of 120 ms and a frequency separation of about one octave. A similar observation was made in A1 of anesthetized cats where the maximally enhanced responses occurred when the first tone was above one octave below or above the subsequently delivered second tone (Brosch and Schreiner, 2000). Brosch et al. (2013) reported that, in the caudomedial field of anesthetized macaque monkey auditory cortex, the likelihood of synchronization between spontaneous firings of a pair of simultaneously recorded multi-unit clusters was highest when the best frequencies of the two clusters were about an octave apart. "
    [Show abstract] [Hide abstract]
    ABSTRACT: A fundamental structure of sounds encountered in the natural environment is the harmonicity. Harmonicity is an essential component of music found in all cultures. It is also a unique feature of vocal communication sounds such as human speech and animal vocalizations. Harmonics in sounds are produced by a variety of acoustic generators and reflectors in the natural environment, including vocal apparatuses of humans and animal species as well as music instruments of many types. We live in an acoustic world full of harmonicity. Given the widespread existence of the harmonicity in many aspects of the hearing environment, it is natural to expect that it be reflected in the evolution and development of the auditory systems of both humans and animals, in particular the auditory cortex. Recent neuroimaging and neurophysiology experiments have identified regions of non-primary auditory cortex in humans and non-human primates that have selective responses to harmonic pitches. Accumulating evidence has also shown that neurons in many regions of the auditory cortex exhibit characteristic responses to harmonically related frequencies beyond the range of pitch. Together, these findings suggest that a fundamental organizational principle of auditory cortex is based on the harmonicity. Such an organization likely plays an important role in music processing by the brain. It may also form the basis of the preference for particular classes of music and voice sounds.
    Full-text · Article · Dec 2013 · Frontiers in Systems Neuroscience
Show more