Intro to Cognitive Neuroscience - Exam 1

Level of emergence

Phenomena at one level depend on, but are different from, the preceding levels. 

• Physics (atoms, subatomic particles) → chemistry (molecules) → biology (cells, proteins) → neuroscience (neurons, brain areas) → cognitive psychology (cognitive processes)

Behaviorism & associationism

• Based on nurture

• Brain driven purely by experience → effectively limitless

  • Operant conditioning: strengthening/weakening of behavior contingent on rewards/punishments

    • Emphasis on animal experiments in controlled environments

• Black box approach: experimentally, could only observe stimulus → response => demonstrating association between environment and observable behavior.

  • Cannot account for many fundamental aspects of the mind = no methods for studying mental structures

Cognitivism

• Nature = nurture => many cognitive abilities are innate

• Complex mental structures: rules, implicit assumptions, concepts, mental models,…

• “Open box”: stimulus → mental representations → response

  • Experiments probe channel capacity * observe patterns of errors = reveal mental structures and processes

Bottom-up

Coarsely structured, simple, low-level representations. → building up from crude elements 

• Example of bottom-up approach: 

  • Barlow’s neuron doctrine: “description of activity of single nerve cell is enough description for functional understanding of nervous system”

    • Ignores levels of emergence 

Top-down

Highly structured, abstract, cognitive representations (most abstract → less abstract) 

• David Marr’s critique of Neuron Doctrine: an algorithm is more likely to be understood by studying the nature of the problem being solved.

  • Focus on tasks being solved & computational constraints on solutions. 

  • Examine how information is represented. 

Recurrent/feedback: interactions between bottom-up & top-down.

Marr’s 3 Levels (Computational theory)

Computational theory: What a process does & why

• Rules and constraints on a process 

• Independent of representation 

Marr’s 3 Levels (Representation & algorithm)

• Representation: Formal system for making information (input/output) explicit. 

  • Can be at the expense of obscuring other information. 

• Algorithm: how to transform input → output

  • Depends on choice of representation. → many algorithms possible for a representation. 

  • Independent of hardware implementation 

Example: Hitting a baseball

• Input: small white circle moves across the retina

  • Inner representations: 

    • Path of white circle on retina: 2-D trajectory 

    • Ball coming toward body: 3-D body centered coordinates 

    • Spinal cord transmit signals to contract/extend specific muscles: muscle coordinates 

Marr’s 3 levels (Hardware implementation)

• How algorithm is physically realized: Same algorithm could be represented in very different technologies. 

Marr’s 3 Levels (Implications for Marr’s levels)

• Top-down approach to cognition: process analyzed a high-level, neglecting lower levels. 

• Cognitive psychology can focus on computational theory, representation & algorithm without worrying about neuroscience. 

  • Cognitive neuroscience bridges the levels → mainly focused on representation & algorithm.

Pseudoscience of Phrenology History

Phrenology became popular in the US in 1830s-1840s → employed as “scientific racism” to justify the enslavement of African Americans & removal of Native Americans from their land. 

• Dr. Charles Caldwell → “African skulls” reveal “tamableness” 

• Dr. Samuel Morton → Native Americans skulls indicate they are “adverse to cultivation, slow in acquiring knowledge” 

Phrenology

Sound claims: 

• Specialization of brain function

• Brain: aggregate of many organs, each having specific psychological faculty. 

Less reputable claims: 

• Mental faculties proportional to size of organ

• 1:1 mapping of brain centers & mental functions. 

Bogus claims: 

• All mental faculties are innate (all nature, no nurture) 

• Organ size measured from the outside by tell-tale bumps on the skull. 

• Character traits as mental faculties: benevolence, friendship, etc.

Phineas Gage

Damage area: 

• Left frontal lobe was almost entirely destroyed

• Damage to left orbitofrontal cortex & medial prefrontal cortex (impulse control, social cognition)

  • To the left of (lateral to) the longitudinal fissure.

• Destroyed all 3 of his left frontal gyri

Behavior: observed personality change. 

• No longer observes social conventions 

• Behaves unethically

• Makes poor personal decisions. 

Broca’s Area

• Paul Broca (France, 1860s) had patients who lost the ability to speak properly but retained language comprehension

  • Damage: lesions to left frontal lobe, 3rd frontal gyrus (inferior left frontal lobe)

    • Anterior/rostral to central fissure and left precentral gyrus.

    • Anterior to Wernicke’s area.

  • Broca’s area located near motor cortex that controls face & mouth movements.

Wernicke’s Area

• Carl Wernicke’s patients (Germany, 1870s) could not comprehend language, but could speak.

  • Damage: lesions in left posterior temporal lobe

    • Posterior to central fissure and central gyri

    • Posterior to Broca’s

  • Wernicke’s area adjacent to primary auditory cortex.

Penfield’s Homunculus

• Body maps in motor cortex & somatosensory cortex (1950s) 

• Neurosurgical stimulation around central sulcus

  • Post-central gyrus (somatosensory cortex): caused touch sensations

  • Pre-central gyrus (motor cortex): caused body movements 

• Takeaway: body is mapped in an orderly fashion.

  • Representation of different areas of the body in the brain = Proportional to number of sensory receptors that we have 

Overview of The Brain

• Cognition implemented in the brain are localized within networks of cortical + subcortical regions 

• Single centers for vision, language, social behavior, etc. do not exist ⇒ networks/systems composed of interconnected brain units. 

  • Each localized unit performs specialized processing, but each unit can participate in many types of behaviors. 

• Mind = concerted operation of multiple systems constituted by above components. 

Hemispheres

• Left cerebral hemisphere → controls inputs from the right side of the body 

• Right cerebral hemisphere → controls inputs from the left side of the body.

• 50k neurons/mm^3

• ~16-25 billion cortical neurons (~20% of brains’ neuron) 

  • ~2.5 mm thick cortex, ~1000 cm^2 cortical surface per hemisphere (2 hemispheres) 

Stains

• Golgi: demonstrates whole cell 

• Nissl: cell bodies 

• Weigert: myelin

Big brain: risky evolutionary gamble

• Large metabolic cost: human (400-500 calories/day → 20-25% of total metabolic needs) 

  • Comparison: monkey 5-7%; chimp & gorilla 8-10%

• Human bottleneck (near extinction event) 

  • Ice age (~900K years ago): 98% of Homo erectus perished

  • For ~120K years, breeding population of humans was only ~1280 total people 

    • 2026: 8.3 billion

Cytoarchitectonics

• Focal architecture of neural cell connections → key basis for subdivision (e.g. Brodmann cortical area subdivisions)

• ~300-330 human brain regions per hemisphere

Example of cytoarchitecture

• Stria of Gennari (Brodmann Area 17; aka striated cortex) 

  • In calcarine sulcus of occipital cortex → ribbon of highly stained cells. 

  • Corresponds to cortical layer IV 

  • Area 17 ends and Area 18 begins where Stria ends. 

    • Function: Primary visual cortex

      • Visual inputs (retina) → lateral geniculate nucleus (LGN) of thalamus → primary visual cortex ⇒ creation of retinotopic map of contralateral visual field.

  • Take-away: can divide cortex into different functional regions based on architecture of cortex. 

Lobes & Major Fissures of Cerebral Hemispheres 

• Frontal lobe.

• Parietal lobe: sensory, attention, etc. 

• Occipital: visual

• Temporal lobe: memory, language

• Also: insular cortex → deep region hidden inside the lateral sulcus (covered by frontal, parietal and temporal lobes)

Sulci

Inward part of groove (less deep) → has two banks of cortical surface.

Fissures

Long, deep grooves on surface of cerebral hemispheres → often a groove without an immediate bottom (deep cracks instead of shallow) 

• Longitudinal fissure: separates left & right hemispheres 

• Central fissure (aka central sulcus): frontal & parietal lobes 

• Lateral fissure (Sylvian fissure): separates parts of frontal & parietal lobe from temporal lobe. 

Gyri

ridge-like elevations (bumps) ⇒ can be found bilaterally. 

• Superior/middle/inferior gyri in frontal, temporal & occipital lobe

• Precentral gyrus: located in the frontal lobe

• Postcentral gyrus: located in the parietal lobe

Planes of Section & Directional Terms

• Horizontal/axial/transverse plane: divides brain into upper & lower planes

• Sagittal (sagittus: parallel) plane: left and right halves

• Coronal plane: front and back halves 

Directional terms

• General: 

  • Superior - inferior: top-bottom/above-below

  • Anterior - posterior: front-behind

  • Medial (middle) - lateral (away from middle) 

• Brain-oriented: 

  • Rostral - caudal: towards the front (nose, beak) - caudal (towards the back, tail) 

  • Dorsal - ventral: top of brain - bottom of brain

• Spinal cord oriented: 

  • Rostral - caudal: towards head - tail

  • Ventral - dorsal: stomach - spinal cord axis (back)

Divisions of the Mature Brain

1. Telencephalon: Largest and highest region of the human brain that contains: 

2. Diencephalon: contains the thalamus and hypothalamus.

3. Mesencephalon: midbrain → most superior part of brainstem. ⇒ tectum (dorsal) and tegmentum (ventral) 

4. Metencephalon: pons (brainstem) and cerebellum

5. Myelencephalon: contains medulla oblongata (brainstem) 

Cerebral Cortex 

The cerebral cortex (part of telencephalon) is a thin 3mm sheet that is folded up.

• 16 billion neurons, 6000 synapses/neuron 

• 100 trillion connections (synapses) between cortical neurons.

• Purpose of folding = enhances surface area; shortens length of cortico-cortical connections (reduces amount of white matter needed) 

Neocortex, Mesocortex & Allocortex

• Neocortex: 6 highly-structured layers ⇒ 90% of cerebral cortex 

• Mesocortex: 6 layers, less well-structured 

  • Paralimbic cortex: cingulate, parahippocampal, insular & orbitofrontal 

• Allocortex: 1-4 layers 

  • Hippocampal complex, primary olfactory cortex

Brain-stem:

• Midbrain: eye movement, vision, hearing, and motor reflexes.

• Pons: Facial movements, balance, sleep

  • Connections to cerebellum: 

    • Superior, middle, and inferior cerebellar peduncles. 

• Medulla oblongata: connection to spinal cords, controls survival functions (breathing, heart rate, etc.) 

Approaches in Cognitive Neuroscience

• Approach A: Brain perturbation approach → highlights causation

  • Brain (perturbation of brain: e.g. neurostimulation, inhibition/excitation of neurons) → Cognition (measure task performance) 

  • Causation requires direct manipulation OR need to address all alternative explanations.

• Neuromonitoring approach → highlights correlation

  • Cognition (manipulate cognitive processes) → brain (measure neural variable)

Tuskegee Syphilis Study (1932-72)

• Unethical study conducted by US Public Health Service and Alabama’s CDC

• Purpose: Observe the natural history of untreated syphilis

• Subjects: 600 African-American Men 

  • 399 with latent syphilis, 201 controls 

  • Impoverished sharecroppers

• Subjects deprived of treatment (penicillin) & deceived (told they would receive free health care from US Federal government)

  • 128 subjects died of syphilis or other study aspects

Belmont Report (1979)

• Respect of persons

  • Obtain informed consent from research subjects

  • Protection of vulnerable populations (minors, pregnant, disabled, old, etc.)

• Beneficience → do no harm

  • Maximizing benefits of research project

  • Minimize risks to research subjects

• Justice

  • Fair selection of research participants for both risks & benefits

Responsibilities of human subjects research ⇒ Belmont

• Approval by IRB 

• Informed consent from patient

  • Patient understands full extent of the experiment, and can contact study coordinator

  • Ensure patient wasn’t coerced by means of threatening or bullying 

  • Support motivation to join or refuse experiment

• Be aware of (side)-effects of clinical trials that were not mentioned & report to proper study coordinator

• Ensure that all patients at least get minimal care needed for their condition. 

HIPAA (Health Insurance Portability & Accountability Act)

• Focuses on privacy of patient records 

  • Does NOT apply to all human subjects research → only to patient medical/dental records. ⇒ therefore not always IRB requirement. 

Animal Research Subjects (Key Assumptions)

• Species possesses similar cognitive function as humans → animal models

• Implemented in similar brain structures

• Behavior: 

  • Bar release, eye movement, overt task

  • Correct response rewarded

  • Catch trials: reveal whether the animal understands the task, rather than just exploiting sth about the experiment. 

Animal Research Subjects (IACUC)

• Institutional animal care & use committee required for any institution using live vertebrate animals in research. 

Controlled Laboratory Experiments (Measure: Accuracy)

Patterns of errors can be informative

• E.g. children overgeneralize grammar rules

  • Conclusion of study: rules learned whole; exceptions learned one at a time. 

• Might also use accuracy to equate to task difficulty. 

Accuracy & channel capacity: e.g. dual task paradigm

• Perform two tasks simultaneously & compare performance with single-task conditions

  • Some tasks compete: visualization from memory & active visual search. 

  • Some tasks are independent: humming a tune from memory & active visual search

• Conclusion: cognitive resources are limited & shareable, but are subdivided into classes. 

Controlled Laboratory Experiments (Measure: RT)

RT measures for representation & transformation: 

• Stimulus is represented in multiple forms (Identical: AA, same letter: Aa, both vowels, both consonants, different categories)

  • Task is easy to do ⇒ therefore RT typically used when accuracy is high.

  • Each transformation takes time → RT used to infer representation & transformation.

• Cognitive subtraction: Measure time for a process to occur by comparing two RTs (one w/ same components & other w/ component OI)

  • E.g. T1 = hit a button when you see light; T2 = hit a button when light is green not red => therefore, T2-T1: time to discriminate between color.

RT: Parallel vs. serial processing

Parallel (many-at-once); serial (one-at-a-time)

Experiment: people given a series of letters to hold → later probed whether letter was previously shown.

• Plotting number of items in memory set & reaction time = linear positive relationship

  • Non-zero slope implies serial processing. 

  • Flat-slope (near-zero) would imply parallel processing => increase in set size does not seem to increase search time (items evaluated at the same time)

RT: Mental rotation

Subjects shown pairs of 3-D objects; varied angle of rotation = asked whether they are the same

• As angle difference increased = reaction time increased linearly 

• Conclusion: objects were mentally rotated in 3-D w/ constant speed → 60 degrees/second 

Ceiling vs. floor effects

• Ceiling effect: compression at top end of response range (e.g. when most people answer questions correctly) 

• Floor effect: compression at low end of response range → observed behavior may be up to chance. 

  • Can impact accuracy, RT, ratings, etc.

Other Cognitive Psychology Research Methods

• Self-reports: e.g. on Likert scales

• Neuropsychological tests: standardized test batteries

  • Example: Patient health questionnaire-9 for depression ⇒ summed score from 0-27 (score >/=10 which indicates further clinical assessment needed)

• Case-studies: single exceptional patients

• Naturalistic observations: observe real-life situations, other cultures

• Computer models

Spatial & temporal limits

• Space: are you able to visualize synapse, dendrite, whole brain? 

  • Finer resolutions (sub 1 mm) = requires more invasive techniques (e.g. eCOG in humans)

• Temporal limits: how quickly can the system capture data or distinguish separate events that happen close together in time.

Resolution vs. field-of-view

• Bottom end of barrier: limit of resolution

  • Resolution: small differences that can be observed

• Top end of barrier/highlight: field-of-view

  • Field-of-view: extent of observations 

• Trade-off: higher resolution → smaller field-of-view

Neuroscience techniques: Animal recordings

• Single unit or multi-unit recordings: ap recorded from neurons

  • Typically limited to one brain area

• Calcium imaging → fiber optic tool records the light emitted from calcium-dependent dyes (during Ca2+ influx from ap) ⇒ get movies of neural activity from neuron population. 

• Optogenetics: causal method.

  • Genetic nucleation put light-activated channels into neurons ⇒ utilize light source (specific colors) to cause excitatory or inhibitory effects in neurons. 

    • Examine effect of certain neuron groups on behavior

Cognitive Neuroscience Techniques: Dissociations & Lesions (Single Dissociation)

Single dissociation: 

• Two groups (e.g. lesion patients & controls) and two tasks. 

• People with specific neuropathology exhibit a deficit in one of the two tasks. 

• Limitation: tasks may not be equally sensitive

Cognitive Neuroscience Techniques: Dissociations & Lesions (Double Dissociation)

• Two (+) groups and two tasks

  • Patients with neuropathology 1

  • Patients with neuropathology 2 

  • Typically normal controls 

• People with different kinds of neuropathologies (different regions) exhibit opposite patterns of deficits. 

  • Tie some connection between region & behavior.

• Limitation: simplifies localization of brain function.

Cognitive Neuroscience Techniques: Structural methods (CT)

Computed tomography (CT scan): reconstitution 3-D image based on variously-positioned x-rays ⇒ whole brain

• Safe and cheap; can localize brain tumor 

• Limitation: fuzzy (poor resolution) 

Cognitive Neuroscience Techniques: Structural methods (MRI)

• Strong static magnetic field align protons

• Small magnetic field gradients at resonant frequency cause spins to tip 

• Spin re-alignment occurs at different rates in different tissue = permitting imaging. 

• Best structural picture of living brain → non-invasive

  • Can localize lesions in living subjects

  • Greater details than CT → soft tissues, nerves, etc.

• Limitation: 

  • Dangerous if the patient has metal (ferrous content) or technology in body. 

  • More expensive

  • Anatomy only → static images.

  • Temporal resolution in minutes range 

Cognitive Neuroscience Techniques: Structural methods (DWI)

Diffusion weighted imaging (DWI/DTI/DSI) 

• Structural technique → uses MR scanner 

• Primarily used to study white matter connectivity 

  • Measures diffusion of water in different directions 

    • Water flows along white matter connections → won’t cross boundaries of fatty myelin = Is myelin intact?

    • Isotropic in gray matter

    • Color represents preferred axis of diffusion if one exists at each VOXEL.

Cognitive Neuroscience: Functional methods (Evoked Brain Potentials)

Electrophysiological basis of EEG, MEG and ERP signals: 

• Measures extracellular current flow.

  • Pyramidal cells have long apical dendrites that broach many cortical layers.

    • Apical dendrites oriented the same way = coordinated electrical signals across population of neurons

• Sensors placed in different regions: Fz (frontal lobe), Cz (center cortex), Oz (occipital), etc.

• Research method/analysis: Cognitive subtraction in neuroimaging

  • Fourier Transform: To find the frequency spectrum, you analyze the time-course signal by systematically matching and subtracting sine waves of various frequencies from it. 

    • With EEG = look at time course signal & do Fourier Transform ⇒ identify the greatest “spike” & frequency. 

Raw EEGs

• Delta (0.1-3 Hz): Deep, dreamless sleep, non-REM sleep, unconscious. 

• Theta (4Hz-7Hz): Intuitive, creative, dream

• Alpha (8-12 Hz): relaxed but conscious brain state. 

• Beta (12-30 Hz): lower range (focused, integrated, thinking) - higher-range (alertness, agitation) 

• Gamma (30 Hz-100 Hz): Motor functions, higher-mental activity

Event-related spectral analysis of EEG: frequency band interactions over time.

Event-related potentials (ERPs)

Average many trials of relative weak EEG signals. 

• Select & average “snippets” standardized to 0 ms time point (stimulus onset) 

• N1/2 = “negativities” ⇒ neuronal spike = NOTE: (-) flipped on graph extracellular environment becomes more negative when ions enter neuron.

• Naming convention: P300: 300 ms after onset; or just order they appear (N1 & 2) 

  • Use ERPs when comparing conditions: visual attention (spatial [attending to specific space/object in space] vs. feature [specific feature/quality] -based) 

    • Found that attended ERPs > than unattended ERPs → different neural signatures for different forms of selection. 

• Can create spatial-temporal maps of ERP activity across time ⇒ using “sensor-space”

MEG

• More spatially precise than EEG, but more expensive. 

  • Must build room for MEG machine = block out all other external electromagnetic signals

  • Must be maintained by cooled liquid gases 

• EEG & MEG dipoles are perpendicular to each other (sine, cosine relationships).

  • MEG gradiometers sit above the skull so can be combined with EEG cap for simultaneous recordings. 

• Advantage: 

  • Skull cannot distort MEG signals like EEG signal. 

• Limitation: 

  • May lose MEG signal due to cortex folding, especially along sulci walls → signal cancelation & dipole orientation. 

Evaluation of Evoked Brain Potential Methods

• Positives:

  • High temporal resolution (milliseconds) of functional methods in humans.

• Limitations: 

  • Can only measure electric signals closer to the scalp (with electrodes) 

  • Fundamental limitations: inverse problem → lead to poor resolution.

    • Forward problem calculates effects or outputs from known causes or inputs ⇒ e.g. if we have a model head w/ known electrical dipoles → we can predict data. 

    • Inverse problem: for all output voltages measured (2D scale) = various/”infinite” possible sources in 3-dimensional layers of human head → hard to pin-point. 

Cognitive Neuroscience Functional Method: Intracranial EEG/eCOG

• Invasive: involves surgically implanted electrodes, e.g. in patients w/ epilepsy

• Correlational

Cognitive Neuroscience Functional Neuroimaging: PET

Positron Emission Tomography (PET): First non-invasive neuroimaging technique available. 

• Measures metabolic brain activity via radioactive labels w/ short half-lives. → injected through IV

• Scanner = passive array of gamma ray sensors 

• Steps: 

  • Wait for protein uptake to go through radioactive decay → release short-lived positron 

  • When positron & electron from gamma ray collide = “annihilate” = detected by detectors. 

  • Look at the time difference between events to obtain a pattern. 

Cognitive Neuroscience Functional Neuroimaging: PET Evaluation

• Limitation: 

  • Difficult to access radio-labels due to short half-lives. 

  • Takes time (long temporal resolution = minutes)

• Main uses: 

  • Can label neurotransmitters, not just glucose or oxygen

    • Example: radio-label for dopamine ⇒ examine uptake in the putamen (in basal ganglia) of Parkinson's disease patients. 

Cognitive Neuroscience Functional Neuroimaging: fMRI

Measures blood-oxygen from brain activity

• Oxyhemoglobin (from lungs) → deoxyhemoglobin (when used) = changes blood magnetic state detectable by fMRI. 

  • BOLD signal: reflects volume & concentration of oxy and deoxy-

    • First small dip = deoxy increase by active neurons

    • Followed by large Oxy increase: compensation for brain activity

• Terms:

  • Voxel: volume element (3-D pixel) 

  • ROI: Region of interest ⇒ defined cluster of voxels that are examined tgt

Cognitive Neuroscience Functional Neuroimaging: fMRI Evaluations

• Pros: 

  • Non-invasive (since hemoglobin are endogenous) ⇒ relies on radio frequency energy

  • Spatial resolution </- 1mm → better than EEG and MEG

  • Wide field of view (brain→column) 

  • Experiment using simultaneous fMRI & sharp electrode recordings on monkey brain demonstrated good correlation w/ local field potentials 

• Cons: 

  • Indirect measure 

  • Modest temporal resolution: delay in signal because neuronal activity must drive metabolic changes = blur

Cognitive Neuroscience Functional Neuroimaging: fMRI Designs

• Block Design fMRI: Due to slow hemodynamic responses, fMRI experiments often block trials together 

  • Blocking: Stimuli/tasks of the same category are presented together in continuous periods (blocks), alternating with rest periods or a different task.

    • Block lengths from 12-40 seconds 

  • ROI analysis: group voxels together into a single time course

• Event-related fMRI: trials can be packed tightly in time

  • Individual stimuli (events) are presented briefly, often in a randomized or mixed order, separated by short time intervals

• Resting-state fMRI: with no explicit task to perform 

  • Observe brain structures that may be correlated with each other => functional connectivity. 

    • Certain frequency bands seem to be critical for communication between brain regions 

Evaluation on Event-related fMRI

• Limitation: 

  • BOLD signals will overlap

  • Can lose some signals due to cramming trials. 

    • Statistical power: Lower per single event, requiring careful timing or varying the space between events to successfully separate overlapping hemodynamic responses

• Advantage: can scramble order of cognitive conditions & determine response due to each after deconvolution analysis

  • Example of research question: Activity during encoding

    • Give participants pictures that they must remember → scan brain 

    • Some time later = give new + old pictures = ask if participants recall image.

    • Sort trials by ones forgotten & remembered = contrast which brain areas were more active during successful/unsuccessful encoding. 

Machine learning & fMRI

MVPA (Multivoxel/multivariate pattern analysis): analyzes distributed patterns of brain activity across multiple voxels to identify networks linked to cognitive functions. 

• Machine learning: Train model on subset of data → test predictions on left-out data. 

Cognitive Neuroscience Functional Neuroimaging: fNIRs

Use light projected through scalp to measure blood oxygenation (through light reflection) 

• Weak-signal w/ a lot of noise (due to abundance of vasculature in the scalp) → Only works for brain signals below skull

  • Has spatial limitation of EEG

• Has temporal limitation of fMRI

• Advantage: can use this technique in groups who can’t do fMRIs (e.g. kids who can’t be still) 

Cognitive Neuroscience Causal, Invasive Methods

• Surgical lesion

• Genetic knockout 

• Optogenetics: 

  • Can stimulate (“on”) or suppress (“off”) using channel rhodopsins (light activated channels; delivered virally) sensitive to different light wavelengths.

  • Typically used in awake or behaving mice. 

  • More focused field of view (neuron); relatively quick resolution. 

• Electrical stimulation 

  • DBS: Deep brain stimulation

• Pharmacology 

Cognitive Neuroscience Causal, Noninvasive Methods: TMS

TMS: transcranial magnetic stimulation

• Time-varying magnetic field induces current flow in tissue (pulse or pulse train) 

  • Weak pulses may stimulate/enhance brain activity

  • Strong pulses briefly disrupt brain activity = transient lesion. 

• Speech arrest examples: 

  • Repetitive TMS: stream of pulses that, depending on duration = effects can persist for seconds or minutes

    • Disruption of Broca’s area = speech arrest (inability to produce speech/say the word)

    • When placed on other hemisphere = no effect → demonstrates lateralization.

Cognitive Neuroscience Causal, Noninvasive Methods: TMS Evaluation

Strengths:

• Excellent temporal resolution compared to other stimulation methods.

Limitations:

• Moderate spatial resolution

• Depth and focality trade off: Weaker magnetic fields can only reach superficial areas of the scalp.

Cognitive Neuroscience Causal, Noninvasive Methods: tDCS

tDCS: transcranial direct current stimulation

• Sends a very weak electrical current through electrodes on the scalp to modify nerve cell activity (e.g. shift resting membrane potential)

• Limitation: non-specific

Cognitive Neuroscience Causal, Noninvasive Methods: tACS

tACS: transcranial alternating current stimulation ⇒ same electrode set up as tDCS, but current oscillates at set frequency to entrain brain’s endogenous rhythm & synchronize neuronal firing.

• TI - temporal interference (tACS, but more focal) 

  • Uses two (or more) electrode pairs, each delivering a high-frequency alternating current → when overlap deep in the brain, they create a low-frequency "beat" that neurons can follow

  • Only overlap zone creates meaningful low-frequency signal = focal, deep stimulation

Cognitive Neuroscience Causal, Noninvasive Methods: tFUS

tFUS: transcranial focused ultrasound ⇒ focused sound waves that converge at specific point 

• Compared to tDCS and tACS, tFUS targets deeper structures at millimeter scale.

• Allow molecules (e.g. helpful drugs) to cross BBB

Types of long-range cortical connections

• Goes through corpus callosum:

  • Homotopic: mirrored on the other hemisphere

  • Heterotropic: connection goes to a different location

• Ipsilateral: within the same hemisphere

Sperry & Gazzaniga: Split Brain Patints

• N=11 underwent surgery (corpus callosotomy) 

• In general, behavioral differences require careful lab testing to observe 

• Non-dominant hemisphere can still process info

  • Deficits observed w/ complex tasks

    • Speech production relies on LH 

    • Complex visual-spatial processing more often relies on RH

• Comparatively, in healthy brains: Most hemispheric asymmetries are subtle (in in-tact brain) ⇒ each side has competence in every cognitive domain (exception of speech production)

Experiment: Patient Joe & Localization of Language Functions

• Subject fixate on central location and object is presented on left & right visual fields. 

  • Normal individual: due to contralateral mapping, left visual field → right visual cortex → moves to left visual cortex → access language system (e.g. Broca’s area) on left hemisphere.) 

  • Split brain information: no pathway for right visual cortex to cross left (EXCEPTION: language performance for stimuli on the right is normal)

    • Interestingly: Patient Joe could not “say” what the object is, but can draw it. 

• Takeaway: localization of certain language functions in left hemisphere for most ppl.

Evidence for LH Dominance of Language: WADA test

• Anesthetize (barbituate in carotid artery) one hemisphere 

• Combine with language study (presentation of object on either visual field) ⇒ is the right hemisphere itself able to identify the object? 

  • Example: amytal injected into left carotid artery = anesthetic LH 

    • Participant presented info on right visual field ⇒ If they are unable to name it = they are LH dominant for language. 

Hemispheric Dominance: Visuospatial Processing & Spatial Attention

Participants align blocks to replicate 2-D picture in 3-D form using different hands. 

• Left-hand performance was better ⇒ Right hemisphere has better visuospatial processing than LH 

Hemi-spatial neglect

Usually follows stroke → Right hemisphere damage to parietal & frontal lobes. 

• Tendency to neglect left visual space (more rare for right side) 

• Not sensory or motor deficit 

• Occurs across sensory modalities

Localization & Pathway: Somatosensory system

Parietal lobe 

• Primary somatosensory cortex (S1): immediately caudal to central sulcus. 

• S2: unimodal association area located ventrally to S1

Touch receptor → ascend dorsal column of spinal cord → 1st synapse at medulla → innervate contralateral thalamus → projects to primary somatosensory cortex (S1)

Types of somatosensory receptors

• Mechanoreceptors: regular (Merkel), light (Meissner’s), deeper pressure (Pacinian) 

• Temperature (ruffini) 

  • Thermoreceptors

• Pain (nociceptors) 

  • Free nerve endings

• Propioception: 

  • Kinesthesia: join movement & position

Topographic organization of somatosensory inputs & Evidence

Somatosensory homunculi: Cortical area (dependent on number of receptors) reflects sensitivity of species 

• Experimental method: find threshold for 2-point discrimination

  • Example: in fingertips, ppl are able to discriminate between pressure on different points due to increased sensitivity 

• Study: looked at somatosensory cortical response (MEG) increases w/ years of training (age of first training) 

  • Finger representation for string players were stronger than control 

  • Strong correlation w/ age of inception = earlier they start, larger the cortical response. → plasticity. 

Phantom limbs & Somatosensory Cortex Plasticity

After amputation of limb, brain’s representation persists. → can still pain/itch (neural response) 

• Possible mechanisms:

  • Voluntary imagination

  • Residual activity in peripheral nerves of amputated limbs.

  • Cortical remapping through plasticity = facial stimulation can drive sensation on phantom hand → due to proximity in homunculus

    • Highlights that “face” region take over “eye” region.

Plasticity in somatosensory cortex: 

• Following amputation of monkey finger (digit 3) = neighboring regions (digit 2 & 4) expand to take over cortical area. 

• Taping finger together can induce some plasticity effects = if the information they provide is redundant, the area kind of shrink/overlap. 

Olfactory system

• Bipolar, chemical receptors at the nasal mucosa

  • There is receptor specificity to chemical types but relies on combinatorial coding 

    • Glomeruli: spherical structures at the surface of olfactory bulb = facilitate many-to-many convergence

• Unique compared to other senses ⇒ No relay through thalamus; pass through into the brain. 

  • Cortex: Localized in the medial, bottom of temporal lobe ⇒ close to thalamus

Gustatory receptors

• 5 types of chemical receptors (sweet, salty, sour, bitter, umami [glutamate]) 

• All receptors in all parts of tongue = sensitivity vary across tongue

Gustatory neural substrates

• Medulla: gustatory nucleus

• Thalamus: ventral posterior medial nucleus

• Cortex

  • Anterior insula

    • Primary gustatory cortex: lateral to olfactory cortex in temporal lobe, tucked in the insula (folded cortex between temporal lobe & frontal lobe)

    • Connects to caudal orbitofrontal cortex ⇒ reward circuitry

Gustatory Senses + Reward: Experiment

Different receptors have different cortical nodes.

• Animal explores two zones, each rely on optogenetic stimulation to activate taste-specific regions in gustatory cortex 

  • Condition 1: Chamber 1 turns on “sweet cortex” 

  • Condition 2: Chamber 2 activates “bitter cortex” 

• Result: animal starts to prefer chamber 1 after training. 

Auditory Receptors

• Hair cells in the cochlea (sensory receptors) in inner ear perceive pressure waves (transformed by bones in middle ear) 

• Frequencies are mapped on the cochlear basilar membrane: 

  • 1-D: High frequency at round window, lower frequencies further up

  • Deafness usually due to destruction of hair cells in cochlea → different receptive fields of frequencies. 

Audition: Sound Localization (Inter-aural time differences)

Difference in time for sound wave to arrive at one ear versus the other.

• Computed in the medial superior olive (MSO). 

• Example: if sound reaches left ear first ⇒ action potential begins traveling toward MSO 

  • Longer path to axon E (left ear leading neuron) 

• Sound reaches right ear later → AP travels towards MSO (shorter path to neuron E; further from A [right ear leading neuron]) 

• Aps converge on MSO neuron that responds most strongly if their arrival is coincident.

Audition: Sound Localization (Inter-aural intensivity/level differences)

Difference in loudness & sound pressure level when sound source is off-center.

• Computed in lateral superior olive (LSO)

  • Each LSO gets two inputs: excitation from ear on its own side & inhibition from opposite ear relayed through inhibitory interneuron (MNTB)

    • Example: 

      • Stronger left-ear signal = strongly excites left LSO; weaker right-ear signal = weak inhibition to left LSO → therefore, net excitation = left LSO signals to higher centers. 

Auditory System: Pathway

• Auditory nerve → cochlear nuclei → superior olivary nucleus → inferior colliculus → medial geniculate nucleus (thalamus) → primary auditory cortex. 

• Tonotopic map of auditory frequency also in primary and secondary belt (auditory cortices) → log map of frequency proportional to # of hair cells. 

• Primary auditory cortex: Superior temporal gyrus of temporal lobe to end of lateral sulcus

Components of the eyes

• Photoreceptors:

  • Rods

  • Humans are trichromats: 3 types of cones (blue, red, green) ⇒ 2.3 million colors. 

    • Red-green color blindness (dichromats): loss/mutation of red/green cones ⇒ struggle to distinguish between these colors. 

• Lens: fine-tune & focus incoming light.

• Fovea: highest concentration of photoreceptors (cones) 

Acuity

• Acuity is determined by distribution of retinal receptors: 

  • Blind spot at the optic disk

  • Rods in higher concentration towards periphery ⇒ less cortex for peripheral vision.

  • More cones concentrated at the fovea. ⇒ more area represented in cortex.

Visual field & Movement of Visual Information

• Half of visual field projects to both retinae: 

  • Right temporal + left nasal → perceives left visual field.

  • Left temporal + right nasal → perceives right visual field 

• Fibers separate at optic chiasm → Each hemisphere (cortex & thalamus) receives input from contralateral visual field.

In the cortex: Moving up pathway = more complex selectivity & larger receptive fields.

• Ventral/temporal pathway: “what” ⇒ recognition

  • V1 → inferior longitudinal fasciculus (white matter fiber bundle) → inferior temporal cortex

  • Conscious pathway

• Dorsal/parietal: “where/how” ⇒ space & guiding motor planning

  • V1 → superior longitudinal fasciculus → posterior parietal cortex

  • Unconscious pathway

Visual cortical receptive field

• Experiment: light bar w/ different orientations are projected on screen. 

  • Neurons fire at specific stimulus orientation 

• In the primary visual cortex: 

  • Ocular dominance columns: stripped bands (mainly in layer IV of V1) that preferentially respond to alternative left-eye & right-eye input.

  • Orientation columns: groups of neurons that respond best to edges/bars of light at specific orientation.  

Ebbinghaus illusion

• Center disks are the same size, but recognition system sees them differently. 

  • Despite appearing different sizes, hand can pick them up without error (different processing pathway for motor system)

• Different perceptual goals require different representations for single stimulus. 

Pathologies: homonymous hemianopia

• Blindness in the opposite visual field in both eyes ⇒ e.g. from occipital lesion in one hemisphere.

• Blindsight: some visual function w/o awareness

  • Damage to primary visual cortex 

  • Experiment: subject w/ hemianopia looks at fixation point & light beam moves across screen + paired w/ tone. Subject is asked to move their eyes to where they think the stimulus is. 

    • Subjects cannot see visual stimuli but their behavior shows they can respond to them (e.g. eye movement) 

      • Demonstrates that bypass to subcortical visual pathway ⇒ superior colliculus to motor regions for eye movement.

Functional specialization

• Color: No motion, colorful display ⇒ activation of V4 (ventral occipito-temporal region)

• Motion: Moving display ⇒ activation of MT (lateral posterior-temporal region)

Other pathologies: visual system

• Achromatopsia: cortical color blindness

  • Damage to ventral occipito-temporal cortex (V4) 

    • Right V4: loss of color perception in left visual field. 

  • Can lose color imagery (cannot imagine what colors look like)

• Akinetopsia: cortical motion blindness

  • Perceptions reported as “still frames” 

  • Patient LH experienced large bilateral damage (area MT)

  • Differences from achromatopsia: 

    • Massive lesion damage

    • Requires both hemispheres 

    • Extremely rare & poorly studied. 

Inverse optics problem: Visual system

Ambiguity in visual perception.

• Problem: The retina gets only a 2-D pattern of luminance, but the brain must recover the 3-D structure of objects. Many scenes could produce the same input, so there is no unique solution.

• Solution: The brain makes unconscious inferences using multiple scene cues plus built-in assumptions about the world.

Visual illusions: Illumination vs. Reflectance

Checker shadow: Luminance = paint shade + lighting, however, eye perceives gray square in light as “lighter” and gray square in darkness/shade as “darker” 

• Assumption: paint changes sharply, lighting changes gradually.

• Soft shadow → discounted, so the shadowed gray looks lighter.