Levelt group

General description research

Our brains are extremely efficient in interpreting highly complex visual information. One theory that has gained increasing recognition in this regard is predictive coding. According to this theory, the brain compares an internal model of the environment with visual inputs it receives. The model is continually updated through prediction errors, which are the disparities between the model and the actual visual input. The Levelt lab is interested in understanding:

• The plasticity mechanisms involved in creating the internal models during development and their updating in adulthood.
• How networks of excitatory and inhibitory neurons that encode feedforward sensory inputs and predictive inputs from higher brain regions are organized.
• How the brain is able to differentiate sensory inputs from the internal model, and how hallucinations may occur when these information streams are not well separated.

Psychiatric symptoms (or disorders) the group is interested in

We have two main psychiatric symptoms that we would like to investigate in collaboration with scientists specialized in psychiatry and/or neurology.

• Hallucinations in the visual system: we want to understand “reality signals” that enable the brain to understand whether information is real, an illusion or imaginary and how altering these signals may result in hallucinations.
• Altered states of awareness: states of reduced consciousness are associated with delta oscillations (3-5 Hz). It has been hypothesized that delta oscillations interfere with the integration of feedforward and top-down inputs required for conscious perception. We developed a mouse model with reduced fast inhibition in the thalamus. In these mice, visual stimulation results in delta oscillations and breakdown of feedforward and top-down inputs. We aim to understand the underlying circuit mechanisms, and possible relationship with humans with reduced states of consciousness, especially those who respond positively to zolpidem or high-frequency stimulation (DBS) of the thalamus.

Technical approaches

We use the mouse visual system to study circuit mechanisms of predictive processing and plasticity. We use approaches to measure the activity of hundreds of individual neurons in awake, behaving mice using two-photon microscopy of calcium signals and Neuropixels electrophysiology. We use gene-targeting approaches to selectively label or modify the neuronal subset we aim to study. Using optogenetics and chemogenetic, or gene-modification approaches, we alter the functions of selective neurons. We combine these techniques with behavioral paradigms that are designed to induce learning and assess perception in mice, and allow us to selectively record responses to feedforward or top-down inputs. We employ powerful data analysis approaches to study the response properties and plasticity of hundreds to thousands of neurons over prolonged periods of times (days-weeks).

Keywords: Predictive processing, hallucinations, psychosis, visual, attention, arousal, dopamine, serotonin, psychedelics, GABA, consciousness
 List of  the matching research domains and/or disorders: Attention, arousal, visual perception