Supramodal processing of discrete quantity information in humans
Human Cognitive and Systems Neuroscience
Final Report Abstract
Previous research in humans and non-human primates has revealed oscillatory neural signatures of how analogue sensory quantities are encoded, maintained in working memory, and integrated into goal-directed comparative judgments. Independently, studies of numerical cognition have deciphered supramodal (i.e. modality-independent) signals in fronto-parietal cortex that encode and maintain discrete quantity information (i.e., numerosity). However, the role of neural oscillations in numerosity processing has for long remained practically unexplored. Here, a series of electroencephalography (EEG) studies was performed to study the role of large-scale cortical oscillations in processing supramodal numerosity. In one set of experiments, numerosity information was presented in form of visual, auditory or somatosensory stimulus trains. Single-trial regression analyses of delayed oscillatory activity showed a prefrontal working memory signature in the beta band (~20-30 Hz), akin to that of analogue stimulus attributes, if the numerosity information was approximated in a quasi-analogue manner. In contrast, to the extent that stimulus trains were processed as individuated sensory events, supramodal modulations of posterior alpha oscillations (8-13Hz) were evident, akin to established oscillatory signatures of working memory load. In particular, comparative analyses within and between the different sensory modalities showed these dissociable working memory mechanisms to arise on a supramodal level of processing, i.e., independent of input modality. A further central objective of the project was methodological, and involved the implementation, application, and validation of convolution-based EEG analysis in the context of sequential numerosity processing. Going beyond conventional single-trial-level analyses, convolution modelling permitted the analysis of ongoing oscillatory activity on a single-event-level, during ongoing, irregular physical stimulation. Complementing recent work in the visuo-spatial domain, these analyses disclosed a critical role of centro-parietal delta-band oscillations (~3Hz) in sequential integration of approximate number: To the extent that the individual events in a stimulus train coincided with the preferred deltaphase, participants tended to over- or underrate the numerosity of the accumulated stimulus train in subsequent comparative judgment. Critically, this type of rhythmic processing did not occur during accumulative encoding of numerosity information per se, but emerged exclusively during formation of a comparative decision. The findings point to a central dissociation of neural dynamics underlying the accumulative encoding and the decisional evaluation, respectively, of sequential numerosity. Together, the project results yield new perspectives on the roles of human alpha-, beta-, and delta-band oscillations in sequential numerosity processing, and advance timely methodological approaches to studying the underlying mechanisms.
Publications
- (2014). Parametric alpha- and beta-band signatures of supramodal numerosity information in human working memory. The Journal of Neuroscience, 34, 4293-4302
Spitzer, B., Fleck, S., & Blankenburg, F.
(See online at https://doi.org/10.1523/JNEUROSCI.4580-13.2014) - Supramodal alpha- and beta-band signatures of numerosity information in human working memory. CNS meeting 2014, Boston, USA
Spitzer, B., Fleck, S., & Blankenburg, F.
- Supramodal rhythmic accumulation of decision-relevant quantity information in humans. OHBM meeting abstracts 2014, Hamburg, Germany
Spitzer, B.
- Rhythmic gain control during supramodal integration of approximate number. NeuroImage, Volume 129, 1 April 2016, Pages 470-479
Spitzer, B., Blankenburg, F., Summerfield, C.
(See online at https://doi.org/10.1016/j.neuroimage.2015.12.024)
