Perception and Sensory Neuropsychology

Sensation

The ability to detect some form of stimulation (raw input)

Perception

The act of giving meaning to a detected sensation (interpretation)

Perception without sensation

Shown by things like hallucinations or phantom limb sensations - proves sensation and perception are separate processes

Three approaches to studying perception

1) Thresholds/psychophysics 2) Sensory neuroscience/neuroimaging 3) Computational models

Thresholds (psychophysics) - question answered

How do we measure perceptual abilities?

Sensory neuroscience/neuroimaging - question answered

What is the biological basis of sensation/perception?

Computational models - question answered

Can maths/computation model perception and how it varies between groups?

Fechner

Founder of psychophysics; often considered the true founder of experimental psychology

Dualism

The philosophical view that the mind exists separately from the physical body/world

Materialism

The philosophical view that only matter exists - mind/consciousness are just results of matter interacting

Psychophysics (definition)

The science of finding quantitative relationships between physical stimuli and subjective (psychological) experience

Two-point threshold

Minimum distance at which two stimuli can be told apart (e.g. two pinpricks on skin)

Detection threshold

Minimum stimulation needed to detect a stimulus, usually defined as being detected 50% of the time

Just Noticeable Difference (JND)

The smallest detectable difference between two stimuli; also called the difference threshold

Method of Constant Stimuli

Many different stimulus intensities are presented one at a time in random order; detection is plotted as % detected vs intensity

Method of Constant Stimuli - key finding

Perception is non-linear - detection increases gradually rather than as a sharp step

Method of Limits

A stimulus is changed step by step (increased or decreased) until the participant's response switches; threshold = average of switch points

Method of Adjustment

Same as method of limits, but the participant (not the experimenter) controls/adjusts the stimulus themselves

Memory hook for the 3 threshold methods

Constant = many random stimuli presented; Limits = experimenter changes it step by step; Adjustment = participant changes it themselves

Weber's Law

The smallest detectable change in a stimulus (JND) is a constant proportion (%) of the original stimulus level, not a fixed amount

Weber's Law - consequence

Larger stimuli need larger JNDs to notice a change; smaller stimuli need smaller JNDs

Weber-Fechner Law example

A 1kg weight needs to change by 0.1kg to notice a difference; a 100kg weight needs to change by 10kg - same 10% proportion, different absolute amounts

The "black box problem" in psychophysics

Psychophysics can control the input (stimulus) and measure the output (subjective experience), but cannot directly observe what happens inside the brain in between

Neurons - firing pattern

Fire in an all-or-none fashion (action potentials/spikes); number of spikes per second reflects how "excited" the neuron is

Action potential pathway

Starts near the cell body, travels down the axon, reaches the axon terminal; involves Na+ and K+ ions moving in and out of the neuron

EEG (Electroencephalography)

Measures electrical activity from populations of neurons via scalp electrodes

ERP (Event-Related Potential)

Electrical activity from a subpopulation of neurons responding to a specific stimulus; requires averaging many EEG recordings

MRI (Magnetic Resonance Imaging)

Uses responses of atoms to strong magnetic fields to image brain structures

BOLD signal (fMRI)

Ratio of oxygenated to deoxygenated haemoglobin; helps localise which brain regions are most active during a task

EEG vs fMRI strengths

EEG = great timing (temporal resolution), weak on location; fMRI/BOLD = great location (spatial resolution), weak on timing

Efficient coding models

Computational models that find predictability in sensory input to encode the world efficiently

Bayesian models

Computational models that make predictions based on prior knowledge about the world

Artificial neural networks

Computer simulations of biological neurons - layers of interconnected units that excite/inhibit each other

Deep Neural Networks (DNNs)

A type of artificial neural network especially good at classifying huge amounts of information into categories; can somewhat predict human perception

Light (definition)

A narrow band of electromagnetic radiation, conceptualised as either a wave or a stream of photons

Cornea

Transparent "window" into the eyeball; light's first entry point

Pupil

Dark opening in the centre of the iris where light enters the eye

Iris

Coloured, muscular part of the eye that controls pupil size (expands/contracts)

Aqueous humour

Watery fluid in the front chamber of the eye

Vitreous humour

Transparent fluid filling the back chamber of the eye

Lens

Structure that focuses the image onto the retina

Retina

Light-sensitive membrane at the back of the eye; contains rods and cones

Optic nerve

Carries the visual signal from the retina to the brain

Order of light's path through the eye

Cornea to Pupil (controlled by Iris) to Lens to Retina to Optic nerve to Brain

Refraction

Bending light so that it focuses properly on the retina (done by the lens)

Accommodation

The lens changing shape to alter its refractive power, allowing focus on near vs far objects

Presbyopia (implied)

Age-related hardening of the lens leads to less effective accommodation, which is why people need reading glasses later in life

Emmetropia

No refractive error; normal vision

Myopia

Light focuses in front of the retina, making distant objects blurry (near-sightedness)

Hyperopia

Light focuses behind the retina, making near objects blurry (far-sightedness)

Astigmatism

Uneven curvature of a refractive surface (usually the cornea), causing generally distorted focus

Memory hook for myopia vs hyperopia

Myopia = focuses too early (in front); Hyperopia = image lands too far back (behind)

Ophthalmoscope

Device doctors use to view the back of the eye (the fundus)

Optic disk

White circle where the optic nerve exits the eye; this is the blind spot because there are no photoreceptors here

Fovea

Central vision spot on the retina, appears brownish, responsible for sharp detailed vision

Photoreceptors

Cells that transduce (convert) light energy into neural energy

Rods

Photoreceptors best suited to low light (night vision); no colour processing, low resolution

Cones

Photoreceptors best suited to strong light (daytime); provide fine detail and colour vision

Distribution of rods and cones

Central vision (fovea) is almost entirely cones; periphery is mostly rods, giving poor peripheral colour vision

Why peripheral vision seems colourful

It is largely an illusion - your brain fills in the gaps since the periphery has few colour-sensing cones

Three types of cones

Short (blue), Medium (green), and Long (red) wavelength-sensitive cones

Visual acuity

The ability to resolve fine detail; correlates directly with cone density, which is highest in the fovea

Photopic conditions

Bright light conditions; cones are the best receptors used

Scotopic conditions

Dim light conditions; rods are the best receptors used

Why colour vision disappears in dim light

Rods dominate in dim light and rods do not process colour

Pupil response to light

Bright light causes pupils to constrict (limit light entering); dark environments cause pupils to dilate (let more light in)

Chromophore

Part of a photoreceptor that captures photons

Opsin

Protein in photoreceptors; rods contain rhodopsin, cones contain 3 opsins (one per wavelength type)

Melanopsin

Found in some photoreceptors; monitors ambient light levels and influences the sleep/wake cycle (circadian rhythm)

Light/dark adaptation (resource depletion model)

More light causes photopigments to be used up faster, leaving fewer available - this desensitisation prevents overstimulation in bright conditions

Graded potentials (photoreceptors)

Signals that vary continuously in amplitude (not all-or-nothing); let the system encode things like a "black surface" as an actual signal

Photoreceptor polarisation during activation

Photoreceptors become hyperpolarised (more negative) when photoactivation starts

Key signalling distinction

Photoreceptors use graded potentials (analog-like); ganglion cells use all-or-nothing action potentials (digital-like)

Vertical retinal pathway

Photoreceptor to Bipolar cell to Ganglion cell (moves signal forward toward the brain)

Diffuse bipolar cells

Receive input from multiple photoreceptors, giving a pooled signal

Midget bipolar cells

Receive input from a single cone, giving a precise signal

Horizontal cells

Run perpendicular to photoreceptors, connect photoreceptors and bipolar cells; responsible for lateral inhibition

Amacrine cells

Connect bipolar cells and ganglion cells horizontally

P ganglion cells

Connect to the parvocellular pathway; receive input from midget bipolar cells; specialise in visual acuity, colour, and object perception; poor temporal resolution but good spatial/detail resolution

M ganglion cells

Connect to the magnocellular pathway; receive input from diffuse bipolar cells; specialise in motion processing; excellent temporal resolution but poor spatial/detail resolution

Memory hook for P vs M pathways

Parvo = Precise details/colour, slow; Magno = Motion, fast but blurry

Receptive field

The region of the retina where stimulation influences a given neuron's (ganglion cell's) firing rate

ON-centre ganglion cells

Excited by light in the centre of their receptive field, inhibited by light in the surround

OFF-centre ganglion cells

Inhibited by light in the centre of their receptive field, excited by light in the surround

What creates centre-surround receptive fields

Horizontal cells, via lateral inhibition

Why centre-surround fields matter

Ganglion cells are most sensitive to contrast between centre and surround (not overall brightness), which helps emphasise edges and boundaries in the real world

Path of visual information from eye to brain

Travels along the optic nerve/tract; crosses over from the left side of fixation to the right side of the brain; both eyes project to both hemispheres

Lateral Geniculate Nucleus (LGN)

A visual relay structure with one in each hemisphere; axons of retinal ganglion cells synapse there

Superior colliculus

A structure involved in eye movements, with one in each hemisphere

Striate cortex

The first visual processing area of the cortex

Ipsilateral

Same side of the body (or brain)

Contralateral

Opposite side of the body (or brain)

LGN structure

Has several layers; not just a simple crossover of left and right eye information

Magnocellular cells (LGN)

Large cells in the bottom two layers (1,2) of the LGN; receive input from M ganglion cells; respond best to coarse detail and moving objects

Parvocellular cells (LGN)

Smaller cells in the top four layers (3-6) of the LGN; receive input from P ganglion cells; respond best to fine detail and stationary objects

Koniocellular cells (LGN)

Very small cells located between the magnocellular and parvocellular layers; their function is less well understood

Striate cortex - other names

Also known as primary visual cortex, area 17, and V1

Naming conventions for V1

Brodmann's map numbers are based on cellular architecture (structure); Visual (V) area numbers are based on physiology; "primary visual cortex" is the cognitive neuroscience term

V1 - major transformation

Circular receptive fields found in the retina and LGN are replaced with elongated "stripe" receptive fields in the cortex ("spots turn to stripes")

V1 layers

The LGN projects mainly to layer 4, with M and P cells going mainly to layer 4C; V1 has approximately 200 million cells, about 100 times more than the LGN that inputs into it