Lecture 13: What is Attention? & Early results

What is Attention? & Early Results

William James on Attention

• William James famously stated, "Everyone knows what attention is" around 1890.
• He also made a noteworthy statement about physics around 1900, suggesting that the field was nearing completion with only more precise measurements left to be made.

Key Concepts and Figures

• Figures: Anne Treisman, Albert Einstein
• Theories: Attenuation Theory, Feature Integration Theory
• Concepts:
  • Context cues
  • Locus of selection
  • Role of distraction for attention
  • Auditory attention
  • Pre-attentive vision
  • Perceptual grouping
  • Illusory conjunctions
  • Parallel vs. Serial search
  • Object Files
  • Parietal role in attention
  • Binding in working memory
  • Statistical perception
  • Scene perception
  • Cross-modal perception
  • Attention and Awareness

Einstein's Theories

• General relativity
• Special relativity
• Photoelectric effect: E=hf
• Mass-energy equivalence: E=mc^2
• Theory of Brownian motion
• Einstein field equations
• Bose–Einstein statistics
• Bose–Einstein condensate
• Gravitational waves
• Cosmological constant
• Unified field theory
• EPR paradox
• Ensemble interpretation

Visual Processing Reminder

• Retinal Cells → RGC & LGN → V1
• V1 cells respond to small dots, orientation, disparity, and some color.
• LOC processes basic shapes.

What Attention Is Not

• Attention is distinct from vigilance, a state of alertness, and mindfulness.

Definition of Visual Attention

• Visual attention is defined as the focusing of the brain's "processing power" on particular regions of the visual input.

Types of Attention

• Overt Attention: Directing the brain’s high-resolution visual processor (the fovea) toward particular parts of a visual scene (i.e., eye movements).
• Covert Attention: Directing the brain’s higher-level processor toward particular regions of a visual scene without moving the eyes.

The Binding Problem

• In the visual cortex, distinct cells are sensitive to different features of the visual input (e.g., orientation-selective, color-selective, motion-selective).
• The binding problem addresses how we perceive whole objects rather than unbound features.

Spatial Colocation and Binding

• Spatial colocation suggests features are bound because they are in the same place. For example, the red color and "T" shape in the upper-left corner are perceived as belonging to the same object because of their location.
• Treisman argues against spatial colocation as the sole mechanism for binding.

Treisman's Early Findings

• Search Tasks: Feature (Disjunction) Search vs. Conjunction Search
• Phenomena: Illusory Conjunctions, Texture segmentation

Treisman's Early Findings: Observation #1

• Visual search is easy if the target differs from non-targets by a simple feature (Feature Search/Disjunction Search).
• Visual search is difficult if the target differs from non-targets by a conjunction of features (Conjunction Search).

Illusory Conjunctions

• Illusory conjunctions involve incorrectly perceiving features from two different objects as one object. For example, perceiving a Red X or a Blue T when those combinations were not actually presented.
• This typically occurs at unattended locations in the visual field.

Treisman's Early Findings: Observation #2

• At attended locations, features are correctly bound together.
• At unattended locations, features are often incorrectly bound together (illusory conjunctions).

Treisman's Early Findings: Observation #3

• Arrangements of colored shapes into textures are easier to discern if different parts of a scene differ in simple features compared to a conjunction of features.
• Textures are repetitive arrangements of similar visual elements that the visual system groups into a coherent region of similarity.

Treisman’s Key Insight

• Some perceptual processes are "attention-free" or "pre-attentive," such as the extraction of simple features like edges. This happens in parallel across the visual field.
• Other perceptual processes require attention, such as binding simple features into coherent object representations. This happens serially at individual locations in the visual field.
• The binding operation at individual locations is the mechanism of attention.
• Processing an object to completion (binding features across different values within and across dimensions into one representation) requires attention.