Biology and neuroscience

AI's ability to process large, complex data sets is transforming biology and neuroscience research.

Further, and perhaps most importantly, at the machine learning/life sciences interface, neuroscience continues to inspire new AI methods and may hold the key to developing generalised AI. Find out how UCL researchers are using AI to solve biology and neuroscience issues below.
 

Biosciences

Using machine learning methods for selective imaging and image analysis to solve problems in biological research that cannot be addressed with current imaging technology.

Machine learning provides exciting opportunities to get the most from large biological datasets and thus increase our understanding of complex processes like ageing. An interdisciplinary project involving three UCL Departments will use fission yeast as a genetic model organism, together with multi-step machine learning, to identify biological processes with fundamental importance for ageing.

Researchers in the Division of Biosciences use AI to understand where species are distributed, why they are distributed in those locations, and how biodiversity responds and feeds into environmental change. Applying modern computational technologies, including Geographic Information Systems, remote sensing, machine learning, and ecological modelling.

As UCL's centre for interdisciplinary research in medical and life sciences, CoMPLEX research is organised across the biological scales that contribute to biological complexity. This includes biomolecular mechanisms, integration of cellular function, physiological and neural systems, and the evolution and dynamics of populations. Projects involve active collaboration between scientists drawn from across UCL.

Cancer

The Chromosomal Instability Research Group investigates the mechanisms by which early somatic mutations and copy-number alterations deregulate cell cycle and the mitotic machinery, giving rise to CIN in non-small-cell lung cancer (NSCLC). The group develop image analysis and use machine learning to implement automated software to reconstruct cell lineages and investigate the effect of specific chromosome segregation errors on cell fate at a single-cell resolution.

Neuroscience

Researchers in the Applied Computational Psychiatry Group are developing computational tools to understand and treat depression, anxiety and addiction. Studies include the use of computational modelling, neuroimaging and behaviour to understand what happens when patients stop taking antidepressants and how this leads to relapse to aid clinical decision making; and the use of computational modelling, neuroimaging and behaviour to understand how decision making and learning contribute to the development, maintenance and relapse in alcohol dependence.

Attention and cognitive control research

Combining research methods spanning neuroimaging (fMRI, EEG, MEG, Spectroscopy), behavioural experiments, psychophysics, and machine learning, the Attention and Cognitive Control Research Group investigates the psychological and neural mechanisms of attention, distraction, visual awareness, executive control, and interaction of emotion, attention and awareness.

Bayesian machine learning

The Gatsby Computational Neuroscience Unit (GNCU) conducts foundational research into the algorithms that underlie perception, action and learning in brains and in machines. GNCU researchers work on a variety of areas of Bayesian statistics, including the use of variational methods to do inference efficiently in complex domains, model selection and non-parametric modelling, novel Markov chain methods, semi-supervised learning and modelling temporal sequences.

Metacognition and computational psychiatry

The Metacognition and Computational Psychiatry Group combine theoretical models with behavioural and cognitive neuroscience approaches including functional and structural MRI, TMS, M/EEG and eye-tracking to understand how the human brain supports self-awareness and metacognition, and how these are altered in mental health disorders.

Researchers in the Sainsbury Wellcome Centre for Neural Circuits and Behaviour (SWC) are combining sensory information with previously learned knowledge to understand brain-wide computations during decision-making in order to develop a mechanistic model that explains the neural dynamics of decision making and choice commitment.

Psychology and language sciences

The Human Behaviour-Change project will build an AI system to continually scan the behaviour change literature, extract key information, and use this to build and update a model of human behaviour.

The Department of Experimental Psychology explores how genes, brains, and experience interact during development to shape how we see and act on the world. Research centres around human learning and decision making. Specific areas of research include understanding the higher-level cognitive processes underlying human capacities of reasoning, judgment and decision making based mainly on behavioural experiments, as well as computational modelling techniques.

The Speech Research Group is producing a machine learning toolkit for the study of the computational modelling of early speech acquisition by infants through real-time interaction with caregivers.