The Virtual Brain Project
The Virtual Brain project aims at developing a computational platform for brain networks that will allow simulating biologically realistic large-scale network dynamics for generic and subject-specific brain anatomy and performing a variety of data analysis.
At the core of the Virtual Brain project lies the nature of the brain as a network. This is illustrated by the insight that interventions in certain brain areas may cause changes, such as functional impairment, in far distant areas (see for instance Corbetta et al 2005).
Brain repair (recovery of function) may depend on the restoration and rebalancing of activity in structurally normal, but functionally impaired nodes of task-relevant networks. The Virtual Brain will exploit these network properties of the brain and provide a platform to systematically study the brain as a large-scale network.
The development of the Virtual Brain will be an on-going process. In the project of INS we wish to lay the foundation for the first step, that is the research group TNG will develop the technical Virtual Brain platform, whereas within this cross-cutting theme we will exploit its capabilities.
The Virtual Brain is a large-scale network model utilizing the primate connectivity matrix between cortex and thalamus.
Brain areas are modelled using neural population models of varying degrees of sophistication. The computational simulations will emulate EEG, MEG and fMRI signals using biologically realistic head models, which will allow a direct quantitative evaluation and comparison with experimental data. All simulations will be performed on a High Performance cluster (click here to access TVB HPC platform), which can be either accessed directly or remotely via a web-based interface. Manipulations of network parameters within the Virtual Brain will allow researchers and clinicians to test out experimental paradigms (such as event-related potentials ERP), clinical hypotheses (such as lesions studies) and therapeutic strategies (such as pharmaceutical interventions targeting individual areas).
The Virtual Brain will give neuroscientists and clinicians access to a sophisticated brain-modelling platform and relevant data analysis algorithms. Most laboratories have neither the computational power, nor the mathematical sophistication, nor the computational neuroscience expertise to perform modelling of this kind. However, it is these laboratories that have the subject and patient specific anatomical and functional data with specific questions and hypotheses about brain function and dysfunction. The Virtual Brain will be designed to test these hypotheses on a scale that could not be performed in a single laboratory.
Our scientific objective is to develop deep insights into the nature and its functioning of the brain as a large-scale network. This shall be achieved through computational simulation, mathematical analysis and experimental testing of the brain’s network character. Our primary goal will be to provide proof of concept for the approach that manipulation of network parameters may be useful for the recovery of brain function. The reverse in fact has been well known to be true for a long time, i.e. the perturbation of a single node can affect the network as a whole (Jackson Br. Med. J. 1884, von Monakow J. Psychol. Neurol. 1911). However, it remains to be demonstrated that the systematic manipulation of network parameters enables the restoration of brain function. If successful, even to the smallest degree, it will have an enormous impact to the field of neuroscience.
The Virtual Brain Project housed at INS has been chosen as a priority area for the Brain Network Recovery Group (Brain NRG http://www.brainnrg.ca). This group is an international consortium funded by the James S McDonnell Foundation and composed of computational, cognitive and clinical neuroscientists dedicated to the applications of neural network theory to understanding of the damaged brain. Brain NRG provides full support of this project and will become one of the key users of the Virtual Brain.
So far, regarding the specific focus on the network character of the brain, the Virtual Brain project is currently unique in the world, though the field is developing quickly.
Creating My Virtual Dream
Public support for science grows when the scientific content is made accessible and personally relevant. We are captivated when a scientific endeavor stimulates both our intellect and our imagination. Space travel and genetic research became far more accessible and interesting to the average person when vivid images of galaxies, nebulae and the DNA helix were used as illustrations. This is particularly salient in neuroscience, where images of the intricate wiring of brain networks reinforces our core desire to understand what makes us tick.
The huge capacity of science to spark the imagination was a driving force for our The Virtual Brain project (TVB, thevirtualbrain.org). Largely as a result of neuroimaging, research on the human brain has advanced enormously in the past two decades, and we now have detailed maps of brain structure and function. Massive amounts of data can be gathered from a single person, and large informatics platforms have evolved to make sense of it all, such as the Human Connectome (www.humanconnectome.org). A big challenge is pulling these data back together to link brain structure and function to mental processes. This is the goal of TVB. We exhibited the key advances in TVB at the Society for Neuroscience (SFN) meeting in 2011 and 2012, where the full software platform was debuted. We received consistent feedback at these events that the artistic feel to the project made it far more enticing and accessible than it might have been with just the computational core. Artistry advancing accessibility and understanding is also quite evident from the video that we developed to convey the concepts behind TVB (www.baycrest.org/thevirtualbrain). Like the images of galaxies, TVB conveys a sense of wonder about what might be possible when we simulate the human brain in a computer model. For example “What would it be like to be inside a brain when it’s dreaming?”
This is what led to the idea for My Virtual Dream (see http://www.myvirtualdream.ca for more details). In this immersive art installation, TVB is depicted as a 60 foot (18 metre) geodesic dome. My Virtual Dream links digitally animated dream sequences and live music performances with twenty participants during 15-minute dream sessions. Participants will use their own brain waves to work with TVB to steer the animations projected on to the giant dome.
Figure . Artist’s rendering of My Virtual Dream installation on the University of Toronto campus in October 2013.
Attendees to the event have two options: (1) To be spectators by staying within the dome’s periphery or (2) Contribute towards creating the overall user experience in the dreamery (the dome’s inner-layer) and communicate with TVB using wireless EEG headbands. A sculpture of TVB will act as the dome’s neural relay – connecting information from participants’ brain waves to the projected animations that illuminate the dome. Groups of up to 20 participants will be able to interact with TVB in the dreamery. (see site plan in “Other Application Materials”)
Dream library: A library of modular video animations will simulate the aspects of dreaming and awareness:
1. δ Delta = slow waves are far reaching in the brain; representing deep sleep and unconsciousness;
2. θ Theta = meditative; representing brain states of REM sleep and dreams;
3. α Alpha = strongest visual imagery;
4. β Beta = focused cognitive processes;
5. γ Gamma = consciousness; representing fast oscillations.
The five libraries will comprise video animations and sound arrangements that are conceptually inspired by the five stages of brain activity across states of awareness.
The Virtual Brain as the DJ: The EEG signals from each participant will be spectrally decomposed and the frequency modulation fed into TVB. The participant’s data will be used by TVB to generate a consensus EEG signal in real-time. Like a disk-jockey, TVB’s consensus signal acts to choose video files and sound recordings from the five dream libraries. The visual manipulations will include morphing, colorization, speed, fading, overlaying and other effects. Similarly, the acoustic impact will comprise analogue synchronous transformations. With each participant contributing their own distinct EEG signal, every audio/visual dream session during the night will be unique.
Figure . Flow diagram of the interactions between participants, TVB, and musicians in building a virtual dream
An electroacoustic experience: Live music will be layered on the video animations. By mapping the participant’s and TVB’s EEG data to MIDI and DSP (digital signal processing) modules, the brain activity will be used to modulate the audio track selected from the dream library. Live musicians will then improvise within this audio scene. Through voice, percussion, trumpet and synthesizer, an audible conversation between participant, TVB and musical performers will be established.
The combination of these actions enables participants to observe the real-time interaction of their own brain waves with an immersive art installation. This installation is a truly unique event at which to disseminate our scientific knowledge and engage the public like never before – through the direct interaction with TVB. Participants will not only see what the brain looks like when it dreams, but also contribute to it. My Virtual Dream will engender the same awe and curiosity about the human brain that drives the original research, and demonstrate the amazing potential when the creative forces of science and art come together.
Please see the following videos to gain some first impression of My Virtual Dream.