Fork the consciousness, or download the project and create your own.

Jakob Garbe

Not stated in the paper

Integrated information Consciousness is intrinsic causal power, measured as integrated information

Jakob Garbe led the construction of DOSE-I, the multimodal biosignal dataset of procedural sedation examined on this site in the DOSE-I analysis. The dataset publishes 78.5 hours of recordings from 171 endoscopy sedation records, with 1,129 annotated consciousness transitions, the moments patients move between conscious and unconscious states, and 7,328 sedation depth labels, alongside preprocessing code. The resource answers a supply problem that shapes the whole measurement field. Indicators of conscious state are easy to propose and hard to validate, because ground truth about state changes is scarce, and clinical archives rarely release it.

The position implicit in the work is that consciousness science advances fastest when its instruments are benchmarked against transitions whose ground truth is annotated by clinicians present at the bedside. Procedural sedation is the best available source of such transitions, since patients drift in and out of consciousness under controlled drugs with continuous monitoring. The dataset’s structure, paired biosignals and two granularities of annotation, lets models learn what a transition looks like across channels rather than in one stream, and makes indicator proposals comparable because they share one annotated ground truth.

The limits are stated plainly in the resource itself, a narrow clinical population, labels built for care rather than for consciousness science, and one path into unconsciousness among many. What the work changes is who can participate. Transition detection research previously required clinical data access. With DOSE-I public, the validation layer beneath every consciousness measure, human or machine, has a shared substrate to be tested against.

Known for. DOSE-I, the first large public dataset of annotated consciousness transitions during procedural sedation, multimodal biosignal benchmarking

Coverage on this site

Related researchers