2. Experimental Methods copy.docx

CRITICAL READING: CORNELL NOTES Experimental Methods Name: Date: 29 July 2023 Section: Lecture 2 Period: Questions/Main Ideas/Vocabulary Notes/Answers/Definitions/Examples/Sentences Sensations Are primitive mental states or experiences induced by sensory stimulation. E.g.: Brightness...

CRITICAL READING: CORNELL NOTES Experimental Methods Name: Date: 29 July 2023 Section: Lecture 2 Period: Questions/Main Ideas/Vocabulary Notes/Answers/Definitions/Examples/Sentences Sensations Are primitive mental states or experiences induced by sensory stimulation. E.g.: Brightness of light. Perceptions Are complex, organised and meaningful experiences of objects or events. E.g.: Recognising a person. We are seeking to understand the relationship between the stimuli in our environment, our neural responses and our perceptual experiences. How do we reconstruct an accurate representation of the 3D visual world around us, from the photons captured by our retina? Cognitive neuroscience: studies the neural machinery. Computational neuroscience: studies the computations of the nervous system. How is sensory information represented in the brain? Cells. Pathways. Attributes. Maps (topographic). Hierarchy of Visual Processing Receptors Photoreceptors (rods and cones) Neurons transmit information Retinal ganglion cells Thalamus (relay station) LGN Cortical processing (occipital lobes) Primary visual cortex (V1) Evidence of functional specialisation: Higher processing “Association areas” The input is modified/refined by the cells at each stage. Visual Processing as Computation Dominant metaphor (currently) for seeking to understand the brain/cognitive processes. Information processing activity of neurons can be represented as mathematical algorithms. Linking Propositions We assume there is a mapping between brain activity and perceptual experiences. A claim that a particular mapping occurs between perceptual and physiological states. Example: The low frequency fall off in the spatial contrast sensitive function is caused by the centre-surround antagonism in receptive fields. These propositions can be falsified through experimentation. It is important to remember that this is the philosophical underpinning of our attempts to understand perception. Experimental Techniques Psychophysics Experimental techniques for measuring the percept associated with a stimulus of a given intensity. Purpose: measuring thresholds Absolute threshold Just noticeable difference Method of Constant Stimuli Across hundred (thousands) of trials, systematically test response of each interval of interest. Distance (width) between test lines. Result: A measure of threshold/sensitivity. The level at which participants are correct on X% of trials. Hyperacuity & Why It Was Important Westheimer & McKee (1977) measured sensitivity to spatial position using psychophysics. Threshold: lower = better performance Found that the ability to discriminate position was better than the width of a photoreceptor = ‘hyperacuity’. Cross referencing results from neurophysiology and psychophysics provided novel insight into sensory processing. How is Hyperacuity Possible? Although the lines were only displaced by 5’’ of arc, the actual lines were much wider. When the second line is displaced by a little, it changes the activity of many photoreceptors, some by up to 25%. Other Psychophysical Methods Method of limits An observer is asked to indicate when a stimulus is present/absent (like a coloured circle). The intensity of a single stimulus changes until a response is made (staircases). Method of adjustment The participant adjusts the stimulus until it is/isn’t detected. Pros of Psychophysics Approaches Valuable behavioural data. Non-invasive. Inexpensive and easy to run. Tight experimental control over all elements. Cons of Psychophysics Approaches Doesn’t measure neural activity. We must make inferences about neural function based on behaviour. Eye Tracking Near infrared light shines from the eye tracker into your eyes. The reflections from your pupil are recorded by cameras in the device, which can be used to calculate where you’re looking. While eye trackers may look old, technological advances mean that some are wearable (glasses, VR systems, etc). Measures (there are many others): First fixation location Fixation duration Regions (areas) of interest – fixation counts Scan paths Provides objective measure of visual experience (e.g. during visual search). Requires minimal training Can test infants or animals Relatively cheap. Can combine with other measures. Gliga et. Al (2009) Eye tracking in 6-month infants revealed a preference to attend to faces compared to objects. Single Cell Recording (Electrophysiology) For sensory information, a single cell codes information present in a small area of space. A ‘receptive field’ is the area over which a cell changes its activity in response to a change in stimulus (e.g. luminance or colour). Recording electrical activity from a single cell using fine insulated wires. Animals Humans V1 – Simple Cells Hubel & Wiesel (1959, 1962): Excitatory and inhibitory regions, arranged side by side. Orientation tuned. Pros: Extremely high spatial/temporal sensitivity. Objective. Cons: Invasive. Doesn’t tell you what is happening elsewhere, though recording multiple cells at once is now possible. Very long, difficult and expensive experiments. Neural Organisation Representation of stimulation is organised in the cortex to mirror stimuli in the world. Each sense keeps a representation of stimuli as a topographic map. Preserves the relative position of stimuli. E.g., Two light dots adjacent in retina will be adjacent in the visual cortex (retinotopic map). Animal Models – Dissection Retinotopic map: The visual pattern was preserved from the retina to V1. Otherwise, dots would be random. Curious: The stimulus has 3 layers, increase in area (1 – 3). Yet, area isn’t proportional in V1. Fovea: cortical magnification Functional Specialisation Clinical Case Studies External injuries (Tatsuji Inouye). Internal injuries: Visuospatial neglect. Often follows right hemisphere strokes. An inability to respond to visual stimuli in the contra-lesional visual field (left). Despite normal vision/motor function. Limitations: Uncontrolled/idiosyncratic Usually no pre-injury assessment Ablation/Lesion Studies A technique in which (small) sections of the brain are damaged or removed. Pierre Flourens (1774 – 1867). Study the subject before surgery. Perform surgery to intentionally damage the brain. Study the subject after surgery. Conclude: Any differences in behaviour are due to the damaged brain region. A cautionary tale: David Ferrier, lesioned the angular gyrus of a tea-loving monkey, resulting in the monkey no longer drinking tea. Later concluded that the monkey is blind but in reality, the angular gyrus is involved in visually guided action. Ungerleider & Mishkin (1982) Two cortical visual systems: Lesions studies and conditioning in Rhesus monkeys. Parietal lobe lesions affected landmark discrimination (LD). Temporal love lesions affected pattern discrimination (PD). Higher Processing – Attributes/Pathways Dorsal stream (‘Where’): LGN to parietal lobe. Motion processing. Spatial processing. Ventral stream (‘What’): LGN to temporal lobe. Object processing. Fine detail. Transcranial Magnetic Stimulation Applies a powerful magnetic pulse to the participant’s head. Temporarily disrupts the transmission of electrical impulses between neurons. Using TMS to study vision: Presented 3 letters to subjects onscreen. TMS bursts over V1 after onset. Bursts at 80 – 100ms impaired accuracy. Suggests that visual information arrives at V1 ~80ms after stim onset. Electroencephalography (EEG) Measures the electrical activity in the brain via electrodes. Pros: Non-invasive Measuring neural activation Good temporal resolution (better than fMRI) Relatively cheap Cons: Poor spatial resolution Depth (origin) Channels Measuring change in 1000’s of neurons Functional Magnetic Resonance Imaging (fMRI) Detects changes in blood oxygenation levels (BOLD imaging). Has been used extensively to infer localisation of function. Pros: Non-invasive Whole brain activity Good spatial resolution > EEG Cons: Only measuring blood flow Temporal resolution Expensive Non-portable Perception Selectivity in fMRI Fusiform face area (FFA) Face selective Lateral occipital (LO) Object-selective Para hippocampal place area (PPA) Place selective Voodoo Correlations? What’s unusual about this fMRI? It’s of a dead Atlantic salmon. The dead salmon was shown images of humans, and asked to decide whether they were feeling socially included or excluded. Significant activation changes occurred. What happened? fMRIs produce a huge amount of data. 130,000 voxels compared over time. Without correcting for multiple comparisons, random noise will be statistically significant. Computational Modelling Using mathematical models (algorithms) to describe, and then eventually predict, behavioural or neural data. E.g.: reaction time, eye movements, or even action potentials in single cells. Assumes that out behaviour, or neural response, is not random – and likely follows a law. By comparing observed data to models, we can see how closely out theory/hypothesis (model) matches reality. Linear Receptive Fields Enroth – Cugell & Pinto (1970): Single cells recording of on-centre RGC of a domestic cat. C) Recording full light on RGC. D) A computer model predicting C, based on A and B. Conclusion: RGC obey simple linear additivity rules. Early computational modelling.