Vision and Perception: Retina to Cortex, Gestalt Principles, and Depth Cues

Vision and Perception: Retina to Cortex — Comprehensive Notes

• Context and safety note
  • Friday special guest lecture: Olivia Scott (Executive Assistant Director of Outreach and Community Engagement for CAF) discussed mental health and suicide risk.
  • Acknowledges that suicide can affect anyone; personal experience shared. Emphasizes the potential value of the lecture for awareness and support, not as a cure for individuals, but as a resource for those affected.
  • Ethical/real-world relevance: importance of mental health literacy and seeking help; sensitivity around traumatic topics in a classroom setting.

Receptive Fields in the Retina

• No single cell sees the entire environment; retina is composed of many receptive fields.
• First major question: What do retinal ganglion cells respond to?
  • Two main types of retinal ganglion cells:
  • On-center receptive field: center responds to light; surround inhibits (on-center, off-surround).
  • Off-center receptive field: center responds to darkness; surrounding light inhibits.
  • Center-surround organization yields contrast sensitivity: bright light in center with dark surround increases firing; light in center plus light in surround reduces net response.
• Demonstration concept (via video clip): changes in light in the center alter action potential firing patterns.
• Conceptual consequence: perception is shaped by context; what you see depends on surrounding light/dark; center-surround antagonism emphasizes edges and contrasts.

Photoreceptors: Rods and Cones; Color Vision

• Photoreceptors transduce light into neural signals; two major types:
  • Rods: high sensitivity, not color-specific; abundant in periphery; essential for low-light vision.
  • Cones: color vision; three types with overlapping spectral sensitivities.
• Three cone types (S, M, L): short, medium, long; commonly mapped to blue, green, red colors.
  • Short-wavelength cones (S): peak sensitivity around $ext{~}420 \, ext{nm}$.
  • Medium-wavelength cones (M): peak around mid-spectrum (roughly green-ish wavelengths).
  • Long-wavelength cones (L): peak around longer wavelengths (red-ish).
• Spectral sensitivity notes:
  • The three cone types are not strict one-to-one; each type has a tuning curve with some overlap.
  • Visual spectrum: $380\,\text{nm} \le \lambda \le \approx 725\,\text{nm}$ (tiny fraction of the full spectrum available).
• Cone distribution:
  • Roughly $85\%\text{--}90\%$ of cones are M and L cones combined, with about $10\%\text{--}15\%$ being S cones.
• Color perception arises from integration across cones:
  • Example: a particular color could produce signals from L and S cones; the brain uses the relative activity across L, M, and S to infer color.
  • The brain’s interpretation depends on the pattern across all three cone types, not on any single cone type alone.
• Color illusions and adaptation:
  • Color afterimages: prolonged stimulation fatigues cones (e.g., center S-cone fatigue leads to afterimage colors opposite the fatigued cone activity).
  • Color constancy: shading and context can shift perceived color even for the same physical color.
• Color vision deficiencies:
  • Most common color blindness: impaired M/L cone function or absence of one cone class; reduces color discrimination (e.g., orange vs green if M/L are misread).
  • When a cone type is missing or malfunctioning, the brain’s combinatorial coding cannot separate some colors that rely on that cone type.
  • Rare cases can involve signaling issues at the photoreceptor level or downstream processing; color perception is context-dependent and not absolute.

From Retina to Cortex: Early Visual Processing and the Pathways

• Retinal ganglion cells project through the optic nerve.
• How visual information crosses:
  • Information from the right visual field initially hits the left hemisphere after crossing at the optic chiasm; information from the left visual field hits the right hemisphere.
  • The optic chiasm is the site of partial crossing; after this, information proceeds to the lateral geniculate nucleus (LGN) of the thalamus for initial cortical processing.
  • From the LGN, signals project to the primary visual cortex (V1; also called area B1 in older nomenclature).
• Primary visual cortex (V1 / area B1): orientation selectivity and edge detectors
  • Orientation-selective cells were famously discovered in V1 via serendipitous experiments in cats (Hubel and Wiesel story). A slide-drum experiment revealed that cells in V1 respond preferentially to edges/lines of specific orientations, not to entire scenes.
  • Classic finding: a single V1 neuron may be maximally responsive to a vertical line; tolerates small deviations from vertical (e.g., 10° off) with decreasing response as the angle deviates.
  • This shows that V1 encodes simple features (edges, orientations) rather than full scenes; perception of complex scenes arises from integrating many simple features across V1 and beyond.
• Visual streams beyond V1:
  • Dorsal stream (the “where/how” pathway): projects to parietal cortex; processes spatial location, motion, and how to interact with objects.
  • Ventral stream (the “what” pathway): projects to temporal cortex; processes object identity and classification (e.g., recognizing a green MSU baseball cap).
  • Information travels along both streams in parallel, enabling simultaneous identification and spatial localization.
• Integration challenge:
  • The brain must combine identity (ventral) with location (dorsal) to form a coherent perception of the world.
  • The fusiform face area (noted in the lecture) shows specialization for faces, with special sensitivity to upright faces and difficulty with inverted faces, illustrating modular processing in higher areas.

Gestalt Principles of Perception

• Core idea: The whole is greater than the sum of its parts; perception is organized by intrinsic grouping rules.
• Simple proximity: elements that are near each other are grouped as a unit.
  • Examples show blue dots forming two or three groups based on spatial proximity.
• Continuation (good continuation): the perceptual system favors continuous lines/paths over abrupt changes; lines tend to be seen as following the smoothest path.
• Similarity: elements that are similar are grouped together (lines of squares vs lines of circles, etc.).
• Figure-ground organization: distinguishing a figure from its background is a fundamental perceptual task; laws can be reversed depending on cues.
  • Figure-ground relationships can flip (white figure on black ground vs black figure on white ground).
• Symmetry and spacing: symmetry and even spacing can influence which regions are perceived as figures versus ground.
• Size as a cue to figure-ground: smaller regions are often seen as figures against a larger background (ground).
• Practical takeaway: perception uses multiple cues to interpret form, size, and boundaries; the brain frequently fills in missing information using these rules.

Adaptive Perception and Depth Perception

• 3D perception from 2D retina is inherently an interpretive process; the brain reconstructs depth using cues.
• The retina is two-dimensional; the world is three-dimensional; depth cues help resolve this gap.
• Two broad classes of cues:
  • Binocular cues (requiring two eyes): mainly binocular disparity (retinal disparity) – slight differences between the two eyes’ views provide depth information.
  • Monocular cues (can be perceived with one eye): several cues still operate well, often used by artists and in static images.
• Binocular disparity (stereopsis): tiny differences between the images in each eye give depth information; the brain computes depth from these disparities.
• Monocular depth cues (examples from the lecture):
  • Occlusion: if one object blocks another, the blocking object is closer.
  • Relative size: if two objects are known to be similar in size, the larger one is perceived as closer.
  • Linear perspective: converging parallel lines indicate depth; the convergence point suggests distance.
  • Texture gradient: textures become finer with distance; denser textures imply depth.
  • Relative height (horizon cue): objects nearer the horizon are typically farther away; height can signal distance.
  • Texture gradient (reiterated): smaller patterns imply greater distance.
  • Motion parallax: as you move, near objects sweep past more quickly than distant ones.
• Practical demonstration concepts:
  • Hippie in the hallway illusion (old demonstration): static 2D cues can produce a perceptual illusion of height or size depending on relative depth cues; demonstrates how context and depth cues influence size perception.
  • Texture and horizon cues illustrate how the brain judges distance even when actual size is constant.
• Inversion and face perception:
  • Inverted faces disrupt face processing; the fusiform face area is specialized for upright faces; when faces are inverted, perception changes due to specialized processing.
• Summary: perception is an active interpretation, not a direct readout of stimulus. The brain uses depth cues to build a 3D interpretation from 2D sensory input. This constructive process can lead to errors or illusions when cues conflict or are ambiguous.

Perception as an Active, Constructive Process

• Seeing 3D in a 2D world requires reconstructive processing; the brain makes educated guesses to fill in gaps.
• These cues optimize processing efficiency but can also yield errors and illusions (e.g., afterimages, color illusions, depth illusions).
• Common reminder: we have a blind spot in our visual field, yet we normally do not notice it because perception fills in gaps using surrounding cues and prior knowledge.
• Takeaway: perception relies on both bottom-up sensory input and top-down contextual interpretation; brain uses expectations, prior experience, and Gestalt principles to make sense of the world.

Review Questions and Exam Prep (Key Concepts and Answers)

• Demand characteristics
  • Question: In an experiment, participants may try to guess the experiment’s purpose and respond to align with the hypothesis; this is demand characteristics.
• A practice item: strength comparison
  • Question: If a value is +0.7 versus -0.7, which is stronger? Answer: They have the same magnitude; strength is the same; sign indicates direction.
• Experimental design example
  • Scenario: A study tests whether psych101 students perform better on exams in cold rooms versus hot rooms, with random assignment to rooms and a common exam.
  • Correct identification: Independent variable is room temperature (hot vs cold).
  • Dependent variable: Exam performance (the result measure).
• Neurophysiology and signaling review
  • Blood-brain barrier (BBB) cell types: The in-class item suggested microglia as responsible; in neuroscience, astrocytes contribute end-feet to the BBB; microglia are immune cells; Schwann cells myelinate PNS. Practical takeaway: be aware of common exam confusions.
  • Myelination cell types: Schwann cells produce myelin in the peripheral nervous system (PNS).
  • Membrane potential and action potentials: at rest, the inside of the neuron is negatively charged relative to the outside; depolarization (positive charge influx) is required to reach threshold and fire an action potential.
  • Direction of ion flow when firing action potentials: sodium influx drives rapid depolarization; potassium efflux follows during repolarization.
  • Resting forces on Na+ at the membrane: both electrostatic attraction (driving Na+ inward when the inside is negative) and concentration gradient (high Na+ outside, low inside) promote inward Na+ flow; thus, both forces act.
  • When discussing the layer of forces during action potential initiation: both electrostatic forces and concentration gradients contribute to the inward Na+ current during depolarization.
  • Monoamine oxidase inhibitors (MAOIs): inhibit enzymatic degradation of monoamines, leading to increased monoamine levels in synapses.
  • Information transfer in the retina: photoreceptors communicate with bipolar cells, which then signal retinal ganglion cells; bipolar cells act as the intermediate transfer cells.
  • Autonomic nervous system cues: parasympathetic activity generally dominates at rest; sympathetic activation increases heart rate and respiration and reallocates resources to body systems needed for “fight or flight.”
• Cellular and systems-level connections
  • Retina-to-brain pathway basics: retina → optic nerve → optic chiasm (partial crossing) → LGN of thalamus → V1 (area B1) in the occipital cortex; beyond that, dorsal and ventral streams process “where/how” and “what” information, respectively.
  • Edge detection and cortical processing: early cortex (V1) emphasizes edges and orientations; higher areas integrate features into object representations; perception arises from multiple parallel streams and Gestalt grouping.
• Real-world relevance and ethics
  • Recognize that mental health topics require sensitivity in presentation; be mindful of triggering content; resources and support mechanisms should be available.
  • Understanding perception helps in fields from design and safety to human-computer interfaces and clinical practice.

Connections to Foundational Principles

• Core sensory-to-cognition pipeline:
  • Receptive fields at retina scale up to complex cortical processing via V1 and beyond.
  • Color vision emerges from three-cone integration; color perception is context-dependent and subject to adaptation.
• Basic neural principles:
  • Action potentials depend on ion gradients and membrane potential dynamics; signaling relies on synapses and neurotransmitter modulation.
  • The blood-brain barrier and glial/neural support cells ensure stable CNS environment; peripheral myelination is handled by Schwann cells.
• Perception is constructive:
  • Gestalt principles, depth cues, and contextual integration show that perception requires interpretation, not just raw data.
• Practical implications:
  • Illusions reveal how the brain uses cues; awareness of perceptual limits can improve education, design, and safety.
• Ethical reflection:
  • The lecture acknowledges the emotional impact of suicide and mental health discussions; emphasizes supportive awareness and accessible help.

Key equations and numerical references (LaTeX):

• Visual spectrum range:
  $380\,\text{nm} \le \lambda \le \approx 725\,\text{nm}$
• Cone sensitivities (summary):
  • S-cones peak: $\lambda \approx 420\,\text{nm}$
  • M-cones peak: in the mid-range (overlap with L-cones)
  • L-cones peak: long wavelengths (overlap with M-cones)
• Cone distribution (typical, from lecture):
  $85{\%}\text{--}90{\%} \text{ of cones are } M \text{ and } L, \quad 10{\%} \text{--} 15{\%} \text{ are } S.$