Highlights

The Highlights

Conclusion of the action

Project results have demonstrated the central role that complexity plays on consciousness. From a theoretical point of view, complexity characterizes consciousness as the equilibrium between integration and information. Furthermore, the dimensions defining different consciousness states as described by Steven Laureys in 2005, which are based on wakefulness and awareness of subjects, have been systematically explored. In this context we realized that the wakefulness dimension is related to the thalamo-cortical connectivity, whereas the awareness dimension corresponds to different degrees of cortico-cortical connectivity. Experimental work has clearly shown that consciousness state categories are not static ones, but that the consciousness states fluctuate depending on the wakefulness degree as a function of the balance between cortico-cortical and cortico-thalamic connectivity. The latter, in particular, appears to act as “consciousness switch”.

From an experimental point of view Luminous has shown that complexity, as measured by an approximation based on the algorithm commonly used to compress data files or on particular spectral features, can be observed in brain signals both spontaneously elicited and after perturbation through non-invasive brain stimulation. This has allowed the implementation of several novel complexity-driven metrics, which have been successfully explored in in-silico simulations as well as in clinical relevant scenarios, i.e., during sleep, in foetal brain activity, anaesthesia, and disorders of consciousness.

Download the final report pdf.

Read Deliverable D1.1: Consciousness: models, metrics & intervention in the electric brain

Check Luminous project on Research gate

Main findings of the third year of the Luminous project

  • We have set-up a theoretical framework of consciousness: Slow-wave activity increases while algorithmic complexity and functional connectivity decrease with loss of consciousness. 
  • We help a networking session entitled as “Human Consciousness meets Artificial Consciousness” in Vienna, and the outcome was that we potentially need a complete model of consciousness that takes into account the following aspects: self-awareness, subjective experience, social realm, biological memory, intelligence and emotions.
  • We developed a realistic computational model of human EEG for consciousness that includes 66 cortical areas and the thalamus, named COALIA.
  • The COALIA model has been successfully shown to reproduce experimental TMS-evoked responses in wakefulness and in sleep that results in similar PCI (Perturbation Complexity Index) values as in real experiments.
  • We found correlation between the Lempel-Ziv complexity values and the SWAS in the resting-state EEG of healthy controls during propofol anaesthesia.
  • We created a new combined EEG and EOG-based auditory communication system through BCI for patients in transition from LIS to cLIS.
  • We have developed a machine learning system for DOC patients that is able to discriminate accurately responders to tDCS from non-responders.
  • We finalised the implementation of the closed-loop tCS protocol for patients with Disorders of Consciousness, and have started the pilots.
  • We have created a model for stimulating the Executive Function Network.
  • We found that combining alpha rTMS stimulation with alpha tACS in the bilateral prefrontal cortex enhances alpha oscillations in the EEG.
  • We found that the strength of the effective connectivity between the frontal and temporal cortices is directly modulated by the consciousnesslevel of the brain.
  • We found early differences of a local and global mismatch in foetal and neonatal subjects that imply that a global rule learning occurs.
  • We validated the simplified version of the famous PCI (Perturbation Complexity Index) computation, and found differences between SEP-induced(Somatosensory Evoked Potentials) and TMS-induced PCI that may imply that the two techniques clearly activate different cortical and subcortical circuits.

Main findings of the second year of the Luminous project

  • Kolmogorov theory of consciousness is now published!
  • We validated the NMM software for producing wake-sleep cycle data through the adaptation of cortico-cortical and thalamo-cortical
  • New interregional communication mechanisms have been incorporated in the NMM: communication through coherence and gating by inhibition.
  • We found higher Lempel-Ziv complexity values on wakefulness compared to propofol anaesthesia in the resting-state EEG of healthy controls.
  • We proposed a multisite-tCS protocol for CLIS patients to increase vigilance and enhance communication performance through BCI.
  • We found that DOC (Disorders of Consciousness) networks are characterised by poor global information processing (network integration) and relatively larger local information processing (network segregation) compared to networks of healthy controls.
  • We implemented a closed-loop tCS protocol for patients with Disorders of Consciousness.
  • We designed a protocol where rTMS pulses induce alpha activity and tACS entrain it to induce lucid dreaming states.
  • We simplified the famous PCI (Perturbation Complexity Index) computation and started working with SEP (Somatosensory Evoked Potentials) as an alternative to TMS-induced PCI.

Main findings of the first year of the Luminous project

The first review meeting has been successfully carried out and here we present our main findings:
  • We modelled the E-field distribution of a patient in Minimally Conscious State (MCS) and a healthy control and we showed that there are differences due to the lesions of the patient, highlighting the need to create personalized models for brain stimulation.
  • We created strategies for tCS targeting networks for MCS and LIS (Locked-In Syndrome) patients, and designed a closed-loop system for stimulating high or low vigilance levels.
  • In a multisite tDCS (transcranial direct currect stimulation) and EEG (electroencephalography) study we demonstrated significant behavioural changes in TBI (Traumatic Brain Injury) In the same study, we estimated complexity from EEG signals under two experimental conditions (multisite tDCS stimulation and placebo), showing that complexity decreases with DOC (Disorders Of Consciousness) right after tDCS stimulation.
  • We showed that it’s feasible to titrate anaesthesia to achieve and maintain SWAS in an individual patient using real-time EEG feedback, and that changes in the thalamo-cortical connectivity at SWAS indicate true perception loss.
  • We trained a NIRS SVM (Near Infrared Spectroscopy Support Vector Machine) classifier on LIS patients in a BCI (Brain-Computer Interface) setting with a priori known true and false sentences and obtained classification accuracy of around 70%. Our next goal is to improve vigilance and BCI performance by stimulating with tCS/EEG.
  • We revealed a global mismatch component as early as 150 ms and latest difference at 500 ms, in a MMN (MisMatch Negativity) local-global paradigm study of healthy adults. Our goal is to extend this to fetuses.
  • We found MMN observed during wakefulness and REM sleep nut not in N1, N2, and N3 sleep.
  • We worked on EEG-based measures of consciousness using a non-invasive perturbational approach in a longitudinal multi-modal evaluation of DOC patients. These measures include the neuronal bistability and the perturbational complexity index. We distinguished between conscious and unconscious state with 100% specificity and sensitivity. We also carried out animal studies and human healthy subjects studies during wakefulness and NREM sleep, as well as TMS (Transcranial Magnetic Stimulation) applied to the primary motor cortex, and showed that the EEG-response to TMS of hand-M1 is larger and accompanied by a late event-related desynchronization.