Neurophysiology and Biophysics for Ai
A.Y. 2024/2025
Learning objectives
This foundational course provides the theoretical basis of neurophysiology needed to address neurocomputation, AI algorithms and their applications to ICT and biomedicine and to provide a perspective for the comparison between natural and artificial intelligence. This includes understanding how neurons and synapses operate inside local microcircuits, how their activity propagates through large-scale networks, and how synaptic and non-synaptic plasticity is generated based on network activity patterns. Moreover, the anatomical organization of large-scale systems will be considered along-with large-scale dynamics and their control of cognitive and emotional states and consciousness. Special attention will be given to biophysical and computational aspects of brain activity at all scales, providing insight on algorithmic properties. Basics of experimental measurements of brain structure and function and of brain biological properties will be provided.
Expected learning outcomes
Students will learn the foundations of central nervous system physiology, at multiple levels: how complex neuronal functions arise from cellular biophysical properties, the rules governing neuronal network activity and plasticity, and how these features combine to generate higher cognitive functions at the macroscale level. The students are expected to learn how brain investigations are performed, from the experimental and computational point of view, and the relevant applications to AI.
Lesson period: Second semester
Assessment methods: Esame
Assessment result: voto verbalizzato in trentesimi
Single course
This course can be attended as a single course.
Course syllabus and organization
Single session
Lesson period
Second semester
Course syllabus
The program is divided into three main sections:
1) Cellular Neurophysiology: a brief introduction of neuroscience and its history; basics of cytology and classification of neurons and glia; the excitable membrane, the generation of the membrane potential; action potentials generation and propagation; introduction to the application of biophysics to represent neuronal membrane physiology. Applications to computational modelling and AI: extended biophysics for neuron modelling; Hodgkin-Huxley theory; Markov chain theory; single neuron models (reconstruction, simulation, validation, optimization).
2) Microcircuit functions: synaptic transmission; excitation and inhibition in neuronal networks (temporal and spatial organization of the E/I balance), synaptic plasticity (short- and long-term plasticity, plasticity rules, Hebbian and non-Hebbian plasticity, heterosynaptic plasticity, spike-timing-dependent plasticity). Applications to computational modelling and AI: extended biophysics for microcircuit modelling; learning rules in network models; artificial neural network theory (simple perceptron, self-organizing networks, hidden layer networks, autoassociative networks, Elman networks); principles of efficient computing with networks; principles of machine learning in engineering and neuroscience.
3) System neurophysiology: gross anatomy of the nervous system; sensory system (e.g., vision, auditory and vestibular system, somatic sensory system); motor system (spinal cord, reflexes, central control of voluntary movement; motor cortex, basal ganglia, cerebellum); hypothalamus, autonomic nervous system and neuromodulators; motivation; emotion; language; sleep; the resting brain and the attention system; consciousness; memory. Applications to computational modelling and AI: extended biophysics for brain modelling; multiscale brain models (bottom-up and top-down principles); virtual brain modelling (TVB); dynamics causal modelling (DCM); neurorobotic controllers; AI and brain functioning for cognitive and sensorimotor control.
For each section, Prof. Mapelli will teach the pure biological part, while Prof. D'Angelo will develop the computational and AI-related applications of the biological concepts.
1) Cellular Neurophysiology: a brief introduction of neuroscience and its history; basics of cytology and classification of neurons and glia; the excitable membrane, the generation of the membrane potential; action potentials generation and propagation; introduction to the application of biophysics to represent neuronal membrane physiology. Applications to computational modelling and AI: extended biophysics for neuron modelling; Hodgkin-Huxley theory; Markov chain theory; single neuron models (reconstruction, simulation, validation, optimization).
2) Microcircuit functions: synaptic transmission; excitation and inhibition in neuronal networks (temporal and spatial organization of the E/I balance), synaptic plasticity (short- and long-term plasticity, plasticity rules, Hebbian and non-Hebbian plasticity, heterosynaptic plasticity, spike-timing-dependent plasticity). Applications to computational modelling and AI: extended biophysics for microcircuit modelling; learning rules in network models; artificial neural network theory (simple perceptron, self-organizing networks, hidden layer networks, autoassociative networks, Elman networks); principles of efficient computing with networks; principles of machine learning in engineering and neuroscience.
3) System neurophysiology: gross anatomy of the nervous system; sensory system (e.g., vision, auditory and vestibular system, somatic sensory system); motor system (spinal cord, reflexes, central control of voluntary movement; motor cortex, basal ganglia, cerebellum); hypothalamus, autonomic nervous system and neuromodulators; motivation; emotion; language; sleep; the resting brain and the attention system; consciousness; memory. Applications to computational modelling and AI: extended biophysics for brain modelling; multiscale brain models (bottom-up and top-down principles); virtual brain modelling (TVB); dynamics causal modelling (DCM); neurorobotic controllers; AI and brain functioning for cognitive and sensorimotor control.
For each section, Prof. Mapelli will teach the pure biological part, while Prof. D'Angelo will develop the computational and AI-related applications of the biological concepts.
Prerequisites for admission
No prerequisite is required
Teaching methods
All course will be based on lectures. The pdf of the slides used during the course will be shared on the Course page on Ariel
Teaching Resources
The slides of the course will be available on Ariel.
The main suggested textbooks are:
- Bear, Connors, Paradiso. Neuroscience, Exploring the Brain. Wolters Kluwer.
- Spitzer. The mind within the net. Bradford books
- Koch and Segev. Methods in neuronal modelling. MIT press
The main suggested textbooks are:
- Bear, Connors, Paradiso. Neuroscience, Exploring the Brain. Wolters Kluwer.
- Spitzer. The mind within the net. Bradford books
- Koch and Segev. Methods in neuronal modelling. MIT press
Assessment methods and Criteria
Written exam with open questions (2/3 of the final grade) and multiple-choice questions (1/3 of the final grade).
BIO/09 - PHYSIOLOGY - University credits: 6
Lessons: 48 hours
Professors:
D'angelo Egidio Ugo, Mapelli Lisa