Paradeisios Alexandros Boulakis
parboulakis.bsky.social
Paradeisios Alexandros Boulakis
@parboulakis.bsky.social
he / him / himbo
researcher at gigaphysiologycognition.uliege.be
mind-blanking, spontaneous thinking, brain-body interactions
dnd, cozy fantasy, freddo espresso and cookies
5) Finally, a Canonical Correlations Analysis between EEG SW-like activity and fMRI connectivity patterns showed that larger SW amplitude and steeper downslope correlated with the hyperconnected brain pattern.
October 16, 2025 at 12:00 PM
4) Here is the catch: the brain organised similar to the “rich” P1 during MB and alertness reports. BUT, during low vigilance, the brain resembled the overconnected P5. Critically, the brain resembled P5 during MB too, but only when participants reported being alert.
October 16, 2025 at 12:00 PM
3) Using K-Means clustering, we found that time-varying functional connectivity could be summarised in five connectivity patterns. These varied from patterns of anticorrelations (P1-3) to patterns with weak synchronization (P4) and overall global synchronization (P5).
October 16, 2025 at 12:00 PM
2) When we examined the fMRI global signal (GS) amplitude, a proxy for cortical arousal, both MB and low vigilance were associated with higher GS amplitude, as in Mortaheb 2022.
October 16, 2025 at 12:00 PM
We found:

1) while MB occurred infrequently, the probability of reporting MB increased when participants reported sleepiness.
October 16, 2025 at 12:00 PM
During the SART, participants were probed about their thoughts (on-task, off-task, or blank) and their alertness level (high or low).
October 16, 2025 at 12:00 PM
We also know that MB is related to a fMRI brain pattern of overall positive connectivity (Mortaheb, 2022).
October 16, 2025 at 12:00 PM
To date, we know that the EEG correlates of MB resemble slow-wave-like (SW) activity, typically observed in NREM sleep, termed “local sleeps” (Andrillon, 2021).
October 16, 2025 at 12:00 PM
6) Finally, a Canonical Correlations Analysis between EEG SW-like activity and fMRI connectivity patterns showed that larger SW amplitude and steeper downslope correlated with the hyperconnected brain pattern.
October 16, 2025 at 11:00 AM
4) Here is the catch: the brain organised similar to the “rich” P1 during MB and alertness reports. BUT, during low vigilance, the brain resembled the hyperconnected P5. Critically, the brain resembled P5 during MB too, but only when participants reported being alert.
October 16, 2025 at 11:00 AM
3) Using K-Means clustering, we found that time-varying functional connectivity could be summarised in five connectivity patterns. These varied from patterns of anticorrelations (P1-3) to patterns with weak synchronization (P4) and overall global synchronization (P5).
October 16, 2025 at 11:00 AM
2) When we examined the fMRI global signal (GS) amplitude, a proxy for cortical arousal, both MB and low vigilance were associated with higher GS amplitude, as in Mortaheb 2022.
October 16, 2025 at 11:00 AM
We found:

1) while MB occured infrequently, the probability of reporting MB increased when participants reported sleepiness.
October 16, 2025 at 11:00 AM
During the SART, participants were probed about their thoughts (on-task, off-task, or blank) and their alertness level (high or low).
October 16, 2025 at 11:00 AM
We also know that MB is related to an fMRI brain pattern of overall positive connectivity (Mortaheb, 2022). (Figure)
October 16, 2025 at 11:00 AM
To date, we know that the EEG correlates of MB resemble those of NREM sleep, aka “local sleeps” (Andrillon, 2021).
October 16, 2025 at 11:00 AM
Feature importance was not arousal-invariant. A classifier trained to classify MB on different arousal conditions considered different features depending on arousal condition.
February 11, 2025 at 1:08 PM
When brain and body features were combined, a balanced random-forest classifier was able to classify MB reports above chance level, and outperformed classifiers trained solely on brain or body features. SHAP values also showed that the model relied on EEG and eye openness to classify MB.
February 11, 2025 at 1:08 PM
Additionally, people’s reports were more likely to transition to MB after sleep deprivation, and less likely to mind wandering.
February 11, 2025 at 1:08 PM
As hypothesized, we found increased occurrences of MB reports after sleep deprivation. We also found that the first 10 trials of the experience-sampling following the high-arousal condition had more MB reports compared to the last 10 trials.
February 11, 2025 at 1:08 PM
We combined experience sampling (n=26) with multimodal signals (EEG, ECG, EDA, Pupil, Resp). Possible thought categories were a) sensations, b) mind wandering and c) MB. Participants also performed the task after intense physical exercise (high-arousal) and after sleep deprivation (low-arousal).
February 11, 2025 at 1:08 PM
What is the utility of including MB in spontaneous thinking? Uncovering the neuronal correlates of thought! MB provides a unique contrast of wakefulness and no content to map the neuronal mechanisms that support content and dynamics.
February 11, 2025 at 12:03 PM
Regarding thought dynamics, we could conceptualize MB either as the transition moment between thoughts or an inability to transition. Since MB is not semantically associated with any thought content, moments of transition could phenomenologically translate as MB.
February 11, 2025 at 12:03 PM
To relate MB to spontaneous thinking, we first review current literature on its neuronal correlates. Across fMRI, EEG, and peripheral physiology, MB is associated with reduced cortical arousal and a brain profile more akin to sleep than wakefulness.
February 11, 2025 at 12:03 PM
Occasionally, people report no thoughts, an experience termed “mind blanking”. MB is typically reported as “no thoughts”, “empty mind” or “attending to nothing”. This phenomenological heterogeneity is further confounded by how MB is communicated to participants.
February 11, 2025 at 12:03 PM