Sensitivity to the temporal structure of rapid sound sequences - An MEG study

Abstract

To probe sensitivity to the time structure of ongoing sound sequences, we measured MEG responses, in human listeners, to the offset of long tone-pip sequences containing various forms of temporal regularity. If listeners learn sequence temporal properties and form expectancies about the arrival time of an upcoming tone, sequence offset should be detectable as soon as an expected tone fails to arrive. Therefore, latencies of offset responses are indicative of the extent to which the temporal pattern has been acquired. In Exp1, sequences were isochronous with tone inter-onset-interval (IOI) set to 75, 125 or 225ms. Exp2 comprised of non-isochronous, temporally regular sequences, comprised of the IOIs above. Exp3 used the same sequences as Exp2 but listeners were required to monitor them for occasional frequency deviants. Analysis of the latency of offset responses revealed that the temporal structure of (even rather simple) regular sequences is not learnt precisely when the sequences are ignored. Pattern coding, supported by a network of temporal, parietal and frontal sources, improved considerably when the signals were made behaviourally pertinent. Thus, contrary to what might be expected in the context of an 'early warning system' framework, learning of temporal structure is not automatic, but affected by the signal's behavioural relevance.

Keywords: Auditory scene analysis; MMN; Magnetoencephalography; Offset response; Omission response; Time perception; entrainment.

Fig. 1

Schematic representation of the stimuli. Experiment 1 (A) consisted of isochronous sequences with one of 3 IOI durations (75, 125 or 225 ms). Experiment 2 contained two types of stimuli: Regular sequences, consisting of a sequential repetition of the 3 IOIs (B), and random sequences (C) where the IOIs were presented in random order. Experiment 3 consisted of the same regular sequences as in Experiment 2 except that these contained occasional frequency deviants which listeners were instructed to detect (D). The sequences were interrupted after a variable duration and brain responses were grouped for analysis according to the IOI expected after sequence interruption. The sequences plotted, save for the random sequence in C, are conditions where the expected IOI, if listeners are able to learn the temporal structure, is 75 ms. The dashed line marks the time of the next expected tone pip.

Fig. 2

Results of Experiment 1. A: Measured brain responses (group RMS), time-locked to the offset of the last tone (0 ms). Offset responses are indicated with arrows. B: Focus on the IOI = 225 ms condition. Right: evoked response. White bars indicate the ultimate 5 tones in the sequence. An offset response (yellow circle) is generated shortly after the expected time of arrival of the missing tone (red dashed line). The response to the last audible tone is reproduced over the response to the missing tone (white dashed curve) to facilitate comparison of response dynamics. Source localization results for that condition are on the left. Plotted are t-maps overlaid on a ch2.nii.gz atlas. Significant clusters for the offset peak > pre-offset (indicated by grey dot) are in superior temporal gyrus (STG) and post-central gyrus (PCG), bilaterally. See also Table 1.

Fig. 3

Comparison of offset latencies obtained in Experiment 2a (two possible IOI permutations) and Experiment 2b (one IOI permutation). No significant difference indicates that increased offset latency in Experiment 2 is not due to pattern variability.

Fig. 4

Results of Experiments 2 and 3 (regular non-isochronous sequences) A: Evoked responses (group RMS) to each of the IOI conditions (and the RAND sequence) in Experiment 2 (dark blue) and Experiment 3 (light blue). 0 ms = offset of the last tone; arrows indicate offset responses (no offset responses were visible in the grand RMS of the IOI = 225 ms condition in Experiment 2). Dashed lines indicate the presentation time of the last 8 tones in the sequence. The next expected (non-arriving) tone is shown in red. Overall the activation patterns indicate significantly delayed offset responses when the sequences are not actively attended. B: Offset response latencies across the three experiments. Top: mean latencies across subjects (corrected by subtracting relevant silent duration from raw RT). Dashed line shows the latency of the CONT conditions, for comparison. IOIs presented in the context of a regularly repeating, non-attended, pattern (Experiment 2) are associated with significantly increased offset latencies, indicating a marked reduction in coding accuracy. Once the sequences are made perceptually pertinent (though listeners were not explicitly attending to temporal structure; Experiment 3) latencies shorten considerably, approaching those measured for isochronous sequences (Experiment 1). Bottom: (raw) latency histograms computed iteratively using Bootstrap (Efron and Tibshirani, 1993). C: Localization results for the IOI = 225 ms condition in Experiment 3. Plotted are t-maps overlaid on a ch2.nii.gz atlas. Significant clusters for the offset peak > pre-offset are in superior temporal gyrus (STG), middle/inferior frontal gyrus (M/IFG) and the parietal lobe (PL) bilaterally encompassing the post-central gyrus (PCG), and the inferior parietal lobule (IPL). See also Table 2.