EEG-representational geometries and psychometric distortions in approximate numerical judgment

Abstract

When judging the average value of sample stimuli (e.g., numbers) people tend to either over- or underweight extreme sample values, depending on task context. In a context of overweighting, recent work has shown that extreme sample values were overly represented also in neural signals, in terms of an anti-compressed geometry of number samples in multivariate electroencephalography (EEG) patterns. Here, we asked whether neural representational geometries may also reflect underweighting of extreme values (i.e., compression) which has been observed behaviorally in a great variety of tasks. We used a simple experimental manipulation (instructions to average a single-stream or to compare dual-streams of samples) to induce compression or anti-compression in behavior when participants judged rapid number sequences. Model-based representational similarity analysis (RSA) replicated the previous finding of neural anti-compression in the dual-stream task, but failed to provide evidence for neural compression in the single-stream task, despite the evidence for compression in behavior. Instead, the results suggested enhanced neural processing of extreme values in either task, regardless of whether extremes were over- or underweighted in subsequent behavioral choice. We further observed more general differences in the neural representation of the sample information between the two tasks. The results suggest enhanced processing of extreme values as the brain's default. Such a default raises new questions about the origin of common psychometric distortions, such as diminishing sensitivity for larger values.

Over- and underweighting of extreme values in decisions from sequential samples.

“…For completeness, we note that in addition to our main finding of adaptive distortions, participants' decisions also showed characteristics that were not encompassed by our model simulations: a "leakage" of sample information over time (i.e., a "recency" bias toward later presented samples), and an overall bias toward larger numbers (e.g., choices were more strongly driven by sample values "9" than "1," although the latter provided equally strong objective evidence). Both of these biases have been reported repeatedly in previous work (Anderson, 1964;Appelhoff et al, 2022aAppelhoff et al, ,2022bCheadle et al, 2014;Hubert-Wallander & Boynton, 2015;Luyckx et al, 2019;Spitzer et al, 2017;Summerfield & Tsetsos, 2015;Weiss & Anderson, 1969;Yashiro et al, 2020), but their precise origin and functional role remain unclear. These biases were not modulated in interpretable ways by our present experimental manipulations, leaving their further investigation to future work.…”

“…Here, we showed that whether people subjectively compress or anticompress numerical values depends on whether they are asked to assess the average value of a single stream or to compare the values of two interleaved streams. Arguably, the latter task is cognitively more effortful, because evaluating a sample's decision value for the comparison requires more cognitive operations (see also Appelhoff et al, 2022b). The pattern of results matches the predictions of our simulations with a psychometric model, which showed that compression yields a performance benefit for noisy observers when tasks are within their processing limit, whereas anticompression improves performance in computationally demanding tasks (i.e., where evaluating a sample properly comes at the cost of missing the decision information in other samples).…”

Optimal Sensor Set for Decoding Motor Imagery from EEG

“…EEG data were first pre-processed by following the standardized preprocessing offered by the PREP pipeline [13], [14]. The data were then filtered with a finite impulse response filter with a highpass frequency of 8 Hz and a lowpass frequency of 30 Hz as these frequencies contain most information related to MI [15].…”