Neural networks for learning and cognitive control
From neurons to knowledge: learning in a dynamic world
We study neuronal and behavioral dynamics in a complex, multi-modal environment as rats discover new knowledge through trial-and-error.
Lines of research:
We are using sophisticated analytic methods to infer internal hypotheses being tested by the rat as it learns the meanings of stimuli (the rules of the task) in a changing environment. We also use psychophysical paradigms to increase attentional focus during learning on one sensory modality and study how this biases the learning process to exclude other sensory modalities.
In this rich behavioral context, we study how noradrenergic neuromodulatory system interactions with the prefrontal cortex can drive exploration of possible rules on one hand and, on the other, focus attention to exploit a newly discovered rule. We record extracellular neuronal activity in the locus coeruleus (LC) and prefrontal cortex (PFC) simultaneously to build a deeper neurophysiological model of how LC ensembles and cell types contribute to cognition.
We develop non-invasive, non-pharmacological methods to activate LC and drive exploration. Based on research showing that monochromatic light in the 460 - 480 nm (blue) range but not other wavelength ranges can affect cognition and indirect evidence that it activates the LC, we are directly recording LC neuron spiking in response to a range of monochromatic wavelengths from 405 nm (purple) to 530 nm (green). We think that monochromatic light can be used to activate the LC and cause a rat to abandon old hypotheses in favor of new ones.
After exploration-based discovery of the rule, rats must engage control processes that inhibit inappropriate responses. We study how the brain mitigates mistaken responses. We record neuronal activity in anterior cingulate cortex (ACC) to study its role in detecting conflicting responses and decipher the signal it sends to the PFC, which increases control, in the form of response inhibition. We also assess LC modulation of this ACC/PFC interaction.
Impact on understanding psychiatric disorders:
Our work translates basic rodent neuroscience discoveries into understanding humans' attentional symptoms experienced by individuals with different psychiatric disorders, including depression, anxiety, attention deficit and hyperactivity disorder (ADHD), schizophrenia, Parkinson's disease, obsessive-compulsive disorder, and Tourette's syndrome. In each of these disorders attention becomes "stuck" on an old stimulus, rule, or particular thoughts, instead of exploring new ideas and new options. Modelling humans' symptoms is difficult in animals; however, our behavioral task is key for translational research because an identical task can be used in rats and humans alike and the observed behaviors are similar.
We aim to develop monochromatic light as an alternative to pharmacological treatments that alleviate impaired attention and learning by boosting LC noradrenaline output (current treatment options: Ritalin/methylphenidate and Straterra/atomoxetine).
Impact on computational models of learning and artificial intelligence:
Our sophisticated behavioral task can refine reinforcement learning algorithms, which are at the core of artificial intelligence (AI). Early behavioral studies in rats and humans formed the basis for reinforcement learning algorithms. These theories may not account for the complex behavioral patterns observed in multi-modal environments, especially when an attentional bias is in play.
- Totah NK, Akil H, Huys QJM, Krystal JH, MacDonald AW, Maia TV, Malenka RC, Pauli WM. (2016). “Complexity and heterogeneity in psychiatry: opportunities for computational approaches.” In: Computational Psychiatry: What Can Theoretical Neuroscience and Psychiatry Teach Each Other? Ed. A.D. Redish and J.A. Gordon. Strungmann Forum Reports, vol. 20, J. Lupp, series editor. Cambridge, MA: MIT Press, in press.
- Totah NK and Plotsky PM. (2005). “Refining the categorical landscape of the DSM: the role of animal models.” In Relational Processes and DSM-V: Neuroscience, Assessment, Prevention and Intervention. Beach SRH, et al. (Eds.) Washington, DC: American Psychiatric Association.
* Denotes shared first authorship / equal contributions
- Totah NK, Logothetis NK, Eschenko, O. (2018). “Shifting perspectives on the noradrenergic system: complex regulation of cortex and cognition over multiple timescales.” (Brain Research - in revision)
- Totah NK, Neves RN, Panzeri S, Logothetis NK, Eschenko, O. (2018). “The locus coeruleus is a complex and differentiated neuromodulatory system.” Neuron, 99:1-14. (Also cited as a previously published version on bioRxiv: https://doi.org/10.1101/109710)
- Totah NK, Logothetis NK, Eschenko O. (2015). “Atomoxetine accelerates attentional set shifting without affecting learning rate in the rat.” Psychopharmacology. 232(20):3697-707.
- * Marzo A, Totah NK*, Neves RM, Logothestis NK, Eschenko O. (2014). “Unilateral electrical stimulation of the rat Locus Coeruleus elicits bilateral response of norepinephrine neurons and sustained activation of the mPFC”, Journal of Neurophysiology, 111(12):2570-88.
- Totah NK, Kim YB, Moghaddam B. (2013). “Distinct prestimulus and poststimulus activation of VTA neurons correlates with stimulus detection”, Journal of Neurophysiology, 110(1):75.
- Bondi CO, Taha AY, Tock JL, Totah NK, Cheon G, Torres G, Rapoport SI, Moghaddam B. (2013). “Adolescent behavior and dopamine availability are uniquely sensitive to dietary omega-3 fatty acid deficiency”, Biological Psychiatry.
- Pehrson AL, Bondi CO, Totah NK, Moghaddam B. (2013). “The influence of NMDA and GABAA receptors and glutamic acid decarboxylase (GAD) activity on attention”, Psychopharmacology. 225 (1): 31.
- Totah NK, Jackson ME, Moghaddam, B. (2013). “Preparatory attention relies on dynamic interactions between prelimbic cortex and anterior cingulate cortex”, Cerebral Cortex. 23(3): 729.
- Totah NK, Kim YB, Homayoun H, Moghaddam B. (2009). “Anterior cingulate neurons represent errors and preparatory attention within the same behavioral sequence”, Journal of Neuroscience. 20: 6418.