Centro-parietal EEG potentials index subjective evidence and confidence during perceptual decision making.

Recent studies suggest that a centro-parietal positivity (CPP) in the EEG signal tracks the absolute (unsigned) strength of accumulated evidence for choices that require the integration of noisy sensory input. Here, we investigated whether the CPP might also reflect the evidence for decisions based on a quantitative comparison between two sequentially presented stimuli (a signed quantity). We recorded EEG while participants decided whether the latter of two vibrotactile frequencies was higher or lower than the former in six variants of this task (n = 116). To account for biases in sequential comparisons, we applied a behavioral model based on Bayesian inference that estimated subjectively perceived frequency differences. Immediately after the second stimulus, parietal ERPs reflected the signed value of subjectively perceived differences and afterwards their absolute value. Strikingly, the modulation by signed difference was evident in trials without any objective evidence for either choice and correlated with choice-selective premotor beta band amplitudes. Modulations by the absolute strength of subjectively perceived evidence - a direct indicator of task difficulty - exhibited all features of statistical decision confidence. Together, our data suggest that parietal EEG signals first index subjective evidence, and later include a measure of confidence in the context of perceptual decision making.

Neuronal signatures of a random-dot motion comparison task.

The study of perceptual decision making has made significant progress owing to major contributions from two experimental paradigms: the sequential vibrotactile frequency comparison task for the somatosensory domain requiring working memory, and the random-dot motion task in the visual domain requiring evidence accumulation over time. On the one hand, electrophysiological recordings in nonhuman primates and humans have identified changes in firing rates and power modulations of beta band oscillations with the vibrotactile frequencies held in working memory, as well as with the mental operation required for decision making. On the other hand, firing rates and centro-parietal potentials were found to increase to a fixed level at the time of responding during the random-dot motion task, possibly reflecting an underlying evidence accumulation mechanism until a decision threshold is met. Here, to bridge these two paradigms, we presented two visual random-dot motion stimuli in a sequential comparison task while recording EEG from human volunteers. We identified a modulation of prefrontal beta band power that scaled with the level of dot motion coherence of the first stimulus during a short retention interval. Furthermore, beta power in premotor areas was modulated by participants' choices approximately 700 ms before responses were given via button press. At the same time, dot motion patches of the second stimulus evoked a pattern of broadband centro-parietal signal build-up till responses were made, whose peak varied with trial difficulty. Hence, we show that known modulations of beta power during working memory and decision making extend from the vibrotactile to the visual domain and provide support for the notion of evidence accumulation as an unconfined decision-making mechanism generalizing over distinct decision types.

Oscillatory EEG signatures of postponed somatosensory decisions.

In recent electroencephalography (EEG) studies, the vibrotactile frequency comparison task has been used to study oscillatory signatures of perceptual decision making in humans, revealing a choice-selective modulation of premotor upper beta band power shortly before decisions were reported. Importantly, these studies focused on decisions that were (1) indicated immediately after stimulus presentation, and (2) for which a direct motor mapping was provided. Here, we investigated whether the putative beta band choice signal also extends to postponed decisions, and how such a decision signal might be influenced by a response mapping that is dissociated from a specific motor command. We recorded EEG data in two separate experiments, both employing the vibrotactile frequency comparison task with delayed decision reports. In the first experiment, delayed choices were associated with a fixed motor mapping, whereas in the second experiment, choices were mapped onto a color code concealing a specific motor response until the end of the delay phase. In between stimulus presentations, as well as after the second stimulus, prefrontal beta band power indexed stimulus information held in working memory. Beta band power also encoded choices during the response delay, notably, in different cortical areas depending on the provided response mapping. In particular, when decisions were associated with a specific motor mapping, choices were represented in premotor cortices, whereas the color mapping resulted in a choice-selective modulation of beta band power in parietal cortices. Together, our findings imply that how a choice is expressed (i.e., the decision consequence) determines where in the cortical sensorimotor hierarchy an according decision signal is processed.

Gamma and Beta Oscillations in Human MEG Encode the Contents of Vibrotactile Working Memory.

Ample evidence suggests that oscillations in the beta band represent quantitative information about somatosensory features during stimulus retention. Visual and auditory working memory (WM) research, on the other hand, has indicated a predominant role of gamma oscillations for active WM processing. Here we reconciled these findings by recording whole-head magnetoencephalography during a vibrotactile frequency comparison task. A Braille stimulator presented healthy subjects with a vibration to the left fingertip that was retained in WM for comparison with a second stimulus presented after a short delay. During this retention interval spectral power in the beta band from the right intraparietal sulcus and inferior frontal gyrus (IFG) monotonically increased with the to-be-remembered vibrotactile frequency. In contrast, induced gamma power showed the inverse of this pattern and decreased with higher stimulus frequency in the right IFG. Together, these results expand the previously established role of beta oscillations for somatosensory WM to the gamma band and give further evidence that quantitative information may be processed in a fronto-parietal network.

Response-Modality-Specific Encoding of Human Choices in Upper Beta Band Oscillations during Vibrotactile Comparisons.

Perceptual decisions based on the comparison of two vibrotactile frequencies have been extensively studied in non-human primates. Recently, we obtained corresponding findings from human oscillatory electroencephalography (EEG) activity in the form of choice-selective modulations of upper beta band amplitude in medial premotor areas. However, the research in non-human primates as well as its human counterpart was so far limited to decisions reported by button presses. Thus, here we investigated whether the observed human beta band modulation is specific to the response modality. We recorded EEG activity from participants who compared two sequentially presented vibrotactile frequencies (f1 and f2), and decided whether f2 > f1 or f2 < f1, by performing a horizontal saccade to either side of a computer screen. Contrasting time-frequency transformed EEG data between both choices revealed that upper beta band amplitude (∼24-32 Hz) was modulated by participants' choices before actual responses were given. In particular, "f2 > f1" choices were always associated with higher beta band amplitude than "f2 < f1" choices, irrespective of whether the choice was correct or not, and independent of the specific association between saccade direction and choice. The observed pattern of beta band modulation was virtually identical to our previous results when participants responded with button presses. In line with an intentional framework of decision making, the most likely sources of the beta band modulation were now, however, located in lateral as compared to medial premotor areas including the frontal eye fields. Hence, we could show that the choice-selective modulation of upper beta band amplitude is on the one hand consistent across different response modalities (i.e., same modulation pattern in similar frequency band), and on the other hand effector specific (i.e., modulation originating from areas involved in planning and executing saccades).

Upper Beta Band Oscillations in Human Premotor Cortex Encode Subjective Choices in a Vibrotactile Comparison Task.

Comparisons of sequentially presented vibrotactile frequencies have been extensively studied using electrophysiological recordings in nonhuman primates. Although neural signatures for working memory aspects of such tasks were recently also identified in human oscillatory EEG activity, homologue correlates of the comparison process are yet unknown. Here, we recorded EEG activity while participants decided which of two sequentially presented vibrotactile stimuli had a higher frequency. Because choices in this type of task are known to be systematically biased by the time-order effect, we applied Bayesian modeling to account for individual choice behavior. Using model-based EEG analysis, we found that upper beta band amplitude (∼20-30 Hz) was modulated by participants' choices. The modulation emerged ∼750 msec before a behavioral response was given and was source-localized to premotor areas.Importantly, the choice-dependent modulation of beta band amplitude was invariant to different motor response mappings and reflected the categorical outcome of the subjective comparison between the two frequencies. Consistently, this pattern was evident for both correct and incorrect trials, indicating that the beta band amplitude mirrors the internal representation of the comparison outcome. Our data complement previous findings in nonhuman primates and corroborate that the beta band activity in premotor areas reflects the categorical outcome of a sensory comparison prior to translation into an effector-specific motor command.