Lecture 1 - 2024 Cognitive Neuroscience Lecture Notes PDF

Lecture 1: Cognitive Neuroscience Linking Mind and Brain: Introduction and Methods of Cognitive Neuroscience  Mind and Brain: An Empirical Example Conten...

Lecture 1: Cognitive Neuroscience Linking Mind and Brain: Introduction and Methods of Cognitive Neuroscience  Mind and Brain: An Empirical Example Content Warning: Video contains footage of a surgical procedure Wilder Penfield merican-Canadian neurosurgeon (1891-1976)  Mind and Brain: An Empirical Example Wilder Penfield’s Direct Homunculus electrical stimulation of cortex Produces "mental" sensations of thinking, perceiving etc., rather than a sense of the brain being stimulated Brain areas dedicated to processing sensory or "a star came down and motor functions for each part of the bod towards my nose", "those fingers and my thumb gave a jump", "I heard the music again; it is like the radio"  Mind and Brain: Cognitive Neuroscience Approach  Cognitive neuroscience aims to provide a brain-based account of cognitive processes (thinking, perceiving, remembering etc.)  Made possible by technological advances in studying the brain that are safer and less crude than, say, Penfield’s method  Historical Foundations Do mental experiences arise in the heart (e.g. Aristotle) or brain (e.g. Plato)? How can a physical substance (brain/body) give rise to mental experiences? = MIND–BODY PROBLEM – Dualism – mind and body are separate substances (e.g. Descartes) – Dual-aspect theory - mind and body are two levels of explanation of the same thing (e.g. like wave–particle duality) – Reductionism - mind eventually explained solely in terms of physical/biological theory These issues still relevant to modern cognitive neuroscience  Historical Foundations (cont.) Early anatomists (16th century) believed ventricles are important Cortex was often schematically drawn (top) or misrepresented like intestines (bottom left) until 18th century Gall and Spurzheim (1810, bottom right) provide an accurate depiction  Phrenology (phren: mind & logos: knowledge, influential from circa 1810 to 1840) Different parts of cortex serve different functions Differences in personality traits manifest in differences in cortical size and bumps on skull Crude division of psychological traits (e.g. "love of animals") and not  Functional Specialization without Phrenology Although phrenology is discredited, the notion that different regions of the brain serve different functions has stood the test of time Termed FUNCTIONAL SPECIALIZATION Modern cognitive neuroscience uses empirical methods to ascertain different functions It does not assume that each region has one function or each function has a discrete location (unlike phrenology), but does assume some degree of specialization of neurons in particular regions  Functional Specialization: Broca’s Observations Patient with left frontal lesion (“Tan”) who could not speak but had otherwise good cognitive abilities Paul Broca Suggested a specialized Louis Victor Leborgne aka “Tan” spent more thanlanguage faculty French physician, anatomis and anthropologist 21 years in the hospital speaking a single word in the brain (1824-1880)  Functional Specialization: After Broca Wernicke later observed a patient with poor speech comprehension, but good production Suggests at least two language faculties in the brain (comprehension v. production) that can be independently affected by brain damage Carl Note that the faculties were inferred from Wernicke empirical observation (unlike phrenological faculties) This inference can be made without necessarily knowing where in the brain German physician, anatomist, they are located psychiatrist and neuropathologis (1848-1905) Minds without Brains: The Computer Metaphor Much of twentieth-century psychology was concerned with observations of behavior, rather than observations of the brain during behavior This led to models of cognition that don’t make direct reference to the brain, e.g. the information-processing models popular from the 1950s onwards The models were inspired by thinking of the mind as a series of routines, like those found in computers  Minds without Brains: The Computer Metaphor (cont.) Thinking of the mind as a computer need not entail a commitment to serial over parallel processes in cognition The Return of the Brain: Cognitive Neuroscience 1970s: structural imaging methods (CT, MRI) enable precise images of the brain (and brain lesions) 1980s: PET adapted to models of cognition developed by psychologists 1985: TMS is first used (a non-invasive, safer equivalent of Penfield’s earlier studies) 1990: level of oxygen in blood used as a measure of cognitive function (the principle behind fMRI)  The Methods of Cognitive Neuroscience Temporal resolution Spatial resolution Invasiveness Adapted from Churchland and Sejnowski (1988). The Methods of Cognitive Neuroscience (cont.)  Challenges to Cognitive Neuroscience (1) It is possible to study the mind without studying the brain (or cognitive theories don’t make predictions about the brain) (2) Functional imaging tells us WHERE and/or WHEN cognition occurs, not HOW (3) Cognitive neuroscience is a new form of phrenology Challenge (1): Studying the Mind without the Brain Analogies often drawn between computer software (mind) and hardware (brain) (e.g. Coltheart, Harley) BUT brain provides causal constraints on the nature of cognition (they are not truly independent) E.g. why is visual word recognition achieved by parallel rather than serial search? Hard to give a purely cognitive answer, but easy to explain in terms of neural processing speed  Challenge (2): WHERE not HOW BUT, do reaction time experiments just tell us WHEN cognition occurs and not HOW? The thing that is measured (e.g. local Henson (2005) blood oxygen v. reaction time) is merely data, and it is a theory (rather than data) that explains HOW Challenge (2): WHERE not HOW (cont.) Adapted from Dehaene et al. (2001).  Consider the question of whether there are visual representations of words that are invariant of case (e.g. radio v. RADIO) Challenge (3): The New Phrenology? Uttal has argued that functional imaging is the new phrenology Stories in the popular press do little to allay these concerns Important to consider computational processes rather than simple localization, and also to consider how brain systems interact. This may avoid a new phrenology  Challenge (3): The New Phrenology? Modern cognitive neuroscience is increasingly more interested in networks rather than modules A set of regions that coordinate together Regions may develop particular specializations because of what they connect to (e.g. reading, faces) New mathematical methods for discovering networks (e.g. ‘graph theory’)  Attendance PIN Electrophysiological Methods  Representations in the Head Mental representation = the sense in which properties of the outside world (e.g. colors, objects, associated memories) are copied/simulated by cognition Neural representation = the way in which properties of the outside world manifest themselves in the neural signal (e.g. different spiking rates for different stimuli) Not straightforward to link these two ideas  personality 6 billion neurons! etween 1,000 to 10,000 onnections per neuron emotions memory language Total number of synaptic connections in the brain is astronomically large Etc. etc. Estimates suggest around 100 trillion to 1,000 trillion (10¹⁴ to 10¹⁵) synapses Total number of grains of sand on Earth, estimated to be around 10¹⁸ grains Total number of stars in the observable universe around 10²² to 10²⁴ stars  Electrical Activity of Neuron Release of neurotransmitters (i.e. glutamate, GABA) that cross the synapse and bind to receptors on the postsynaptic neuron  From Neuron to Thought Neurons produce action potentials The height (amplitude) of the action potential is fixed BUT the number of action potentials per second (called ‘spiking rate’) varies Some neurons produce more spikes for some stimuli than for others – i.e. they have a selective response  Stimulus A Stimulus B Stimulus C Stimulus D Neuron 1 Neuron 2 Neuron 3  Stimulus A Stimulus B Stimulus C Stimulus D Neuron 1 Neuron 2 Neuron 3  Two Main Electrophysiological Techniques (1) Single-cell recordings – Electrode(s) placed in or near a neuron (invasive) – Measure number of action potentials per second (2) Event-related potentials (ERP) or Event-related fields (ERF) – Electrode(s) placed on the skull or magnetic sensors placed around the head – Measures summed electrical potentials or associated changes in the magnetic field  Single-Cell Recordings Very small electrode implanted into axon (intracellular) or outside axon membrane (extracellular) Records neural activity (but doesn’t stimulate it, e.g. + like True Penfield) neuronal activity and great spatial and temporal resolution - Invasive and records only from a specific region  Grandmother Cells? A grandmother cell hypothetically responds to (or represents) only one stimulus The ‘Jennifer Aniston’ Neuron? Quiroga et al. (2005) The ‘Halle Berry’ Neuron? Quiroga et al. (2005) The ‘Sydney Opera/Baha’i Temple’ Neuron? Quiroga et al. (2005) Discussion of Quiroga et al. (2013) Does a grandmother cell mean that a single person is represented by a single cell? Do the cells discovered by Quiroga just respond to visual stimuli? Are their responses learned or innate?  Event-Related Potentials (ERPs) Based on EEG (electroencephalography) recordings EEG signal is averaged over many events (to reduce effects of random neural firing) and synchronized to some aspect of the event (e.g. onset of stimulus, pressing a button) Electrodes record a series of positive and negative peaks Timing and amplitude of the peaks is related to different aspects of the stimulus and task (e.g. consider face recognition)  Event-Related Fields (ERFs) Based on MEG (magnetoencephalography) recordings Same process of averaging MEG signal over many events Advantages over EEG – Cleaner data (less signal attenuation, less artifacts, etc.) – Better spatial resolution Disadvantages over EEG – Much more costly – Current technology requires participants to keep their head very still, but watch out for OPMs…  EEG/ MEG topographic maps of evoked data https://mne.tools Advantages and Disadvantages of ERP ERP signal is directly related to neural activity and this electrical activity is conducted instantaneously to the scalp Therefore, ERP has an excellent temporal resolution The ERP signal is derived from different sources in the brain and it is not possible to infer exactly where these sources are from the scalp  Event-Related Potentials (ERPs) (cont.) EEG signal is averaged over many events... From Kolb and Whishaw (2002). Copyright © 2002 by Worth Publishers. Used with permission. Event-Related Potentials (ERPs) (cont.)  Using ERP to Study Face Recognition Different ERP peaks associated with different aspects of face processing Using ERP to Study Face Recognition (cont.) The N170 is relatively specialized for faces Objects (no N170) Human Face Dog Face (N170) (N170) Using ERP to Study Face Recognition (cont.) One can then ask questions about the nature of the N170... – Is it limited to human faces - no – Is it greater for famous faces - no – Is it found for smiley faces  - yes From Bentin et al. (2002). Using ERP to Study Face Recognition (cont.) Seeing one face may make a person think of another person too Does this happen at semantic level or perceptual level? ERP is helpful here, as these stages are broadly separated in time Evidence suggests it happens at late stages (ERP changes after 300msec) Schweinberger et al., 199  Next lecture.... Brain imaging (fMRI) Patient-based neuropsychology Brain stimulation (TMS, tES)