Cognitive Sciences

Front: What are the two main types of brain cells?
Back: Glia and Neurons

Front: What is the function of glial cells?
Back: Support cells that play structural and metabolic roles in maintaining the brain. They:

• Maintain the biophysical environment for nerve function

• Provide armature

• Manage blood compounds' access to neurons

• Serve as natural insulation

Front: What is the function of neurons?
Back: Neurons perform computations and form the foundation of mental function.

Front: What is the cell body of a neuron?
Back: The large center of the cell that contains all machinery needed to keep the cell alive, processes sugars and oxygen, and contains DNA.

Front: What are dendrites?
Back: Inputs to a nerve cell that allow for the mathematical integration and analysis of signals from other cells.

Front: What is the function of an axon?
Back:

• Acts as the output wire for the neuron

• Broadcasts outputs of dendritic computations to other neurons

• The tips of axons make physical contact with dendrites of other neurons

Front: What is the nerve terminal (axon terminal)?
Back: The contact point between neurons, specialized to maximize computational flexibility.

Front: What is a synapse?
Back: The junction between an axon terminal and a dendrite that allows a receiving neuron to perform calculations on received signals.

Front: What are ions?
Back: Atoms or molecules that have a net electric charge.

Front: What is electric charge?
Back: A basic property of matter carried by some elementary particles that governs how particles are affected by an electric or magnetic field.

Front: What is the neuronal membrane?
Back: The neuronal membrane restricts the flow of chemicals in and out of the cell and contains channels that allow permeability to water-based chemicals.

Front: What is diffusion in the movement of ions?
Back:

• Ions dissolved in water are in constant movement

• Ions distribute evenly in a solution from high to low concentration regions

Front: How does electricity relate to the movement of ions?
Back:

• Opposite charges attract, like charges repel

• Electric potential (voltage): Force exerted on a charged particle, reflecting the difference in charge between negatively and positively charged terminals

• Membrane potential: Voltage across the neuronal membrane at any moment

Front: What is resting membrane potential?
Back:

• Restriction of flow through the membrane creates a stable equilibrium

• Diffusion: High concentration of sodium outside the neuron pulls sodium inside

• Electricity: Positive charge of the cell pushes excessive sodium outside

• Resting potential: The equilibrium point

• Opening the channels increases diffusive force, shifting to a higher voltage inside the cell

Front: What are neurotransmitters?
Back: Chemicals that temporarily open ion channels on dendrites, briefly changing the electrical voltage across each neuron's membrane.

Front: What is an action potential?
Back:

• Ion channels in axons are voltage-gated

• Channels open when the voltage near them exceeds a fixed threshold, further increasing voltage

• A wave of action potential propagates down the axon to the terminal

• Frequency of openings can communicate continuous voltage levels rather than binary signals

Front: What is the firing rate of a neuron?
Back: The rate at which action potentials are generated, which is roughly a linear function of dendritic voltage.

Front: How does the brain encode negative signals and limited precision?
Back:

• The brain encodes separate positive and negative segments

• The range of firing rate is limited to 50Hz-zero, which restricts precision

• This limitation can be overcome by involving more than one neuron

Front: What is the metabolic cost of information transmission?
Back:

• Generation of action potentials, neurotransmitter release, and dendritic equilibrium maintenance are metabolically costly

• 20% of oxygen and sugar in an adult brain, and 50% in children, is consumed for these processes

Front: What is information transmission without spikes?
Back:

• Light enters the eye, stimulating the retina and exciting a neuron

• The first two layers of neurons are in constant communication, meaning no spike is needed

Front: Why does the brain use spikes?
Back:

• Accuracy: Reacts to changes in less than a millisecond

• Speed: Ensures rapid communication

• Distance: Signals can travel along meters of axon

Front: What are the main subdivisions of the brain?
Back:

• Telencephalon (~ cerebrum): Includes the cerebral cortex and basal ganglia

• Metencephalon (~ cerebellum)

• Brain stem

Front: What is the function of the basal ganglia?
Back:

• Striatum encodes values of options

• Receives inputs from the frontal cortex and sends outputs to other basal ganglia regions

• Particular interest: dopamine neurons, which encode reward-prediction error signals

Front: What is the cerebral cortex?
Back:

• A large, 6-layered sheet, with each layer serving a function

• Grey matter: Cell bodies

• White matter: Runs of axons serving connectivity

• Brodmann areas indicate functional subdivisions

Front: What is the amygdala responsible for?
Back:

• Part of the telencephalon

• Super old structure responsible for fear

• Receives sensory inputs and sends output to the hypothalamus for somatic responses

Front: What are the main types of neuroscience research methods?
Back:

• Measurement techniques: Correlational, observe brain function changes

• Manipulation techniques: Causal, examine how perturbations change behavior

Front: What factors determine the choice of research method?
Back:

• Temporal resolution: Frequency of measurements

• Spatial resolution: Ability to distinguish adjacent regions

• Invasiveness

Front: What is the absolute threshold?
Back: The minimum intensity of a stimulus that can be reliably detected.

Front: What is the difference threshold?
Back: The smallest difference between two stimuli that can be reliably detected.

Front: What is Weber's law?
Back:

• The Just Noticeable Difference (JND) follows a constant ratio between the change and the base.

• This ratio can describe the difference threshold.

• Weber fraction formula: ΔR = cR or ΔR/R = c

• A modified version includes sensory noise (a): ΔR = c(R + a), where a is the amount of sensory noise that exists when R = 0.

Front: What are the assumptions of Fechner's law?
Back:

1. Weber's law is correct.

2. JNDs are equal psychological increments in sensation magnitude regardless of ΔR size.

• Formula: S = k log R

• S = Magnitude of psychological sensation

• R = Intensity of physical stimulus

Front: What is Steven's law?
Back:

• Describes sensation magnitude estimation based on stimulus intensity changes.

• Records S by asking subjects to rate perception (e.g., from 1 to 10).

• Formula: S = aR^b

• b values differ for different stimuli:

  • Brightness: 0.3

  • Lengths: 1

  • Electric shock: 3.5

Front: What is the perception of numerosity?
Back:

• Counting up to 3 is easy, but beyond that, it becomes difficult.

• Subitization: Fast counting without actually counting.

Front: What factors influence numerosity perception?
Back:

• Overestimation: We tend to overestimate numerosity if objects are regularly spread out on a page.

• Distance effects: Two distant numerosities are easier to distinguish.

• Magnitude effects: For equal distance, we struggle more to distinguish between larger numerosities.

• Formula: (N₁ - N₂) / (N₁ + N₂)

Front: What is the mental compression of large numbers?
Back:

• We tend to compress large numbers into a log-like scale rather than a linear one.

Front: How do we perceive the number line?
Back:

• We associate larger numbers with the right side.

• This perception does not depend on literacy.

• Early in life, numbers are perceived in a logarithmic manner, but after around 4th grade, this perception transitions to a linear scale.

Front: What is Bayes' rule?
Back:

• Formula: P(A | B) = (P(B | A) * P(A)) / P(B)

• Extended version: P(A | B) = (P(B | A) * P(A)) / [(P(B | A) * P(A)) + (P(B | ¬A) * P(¬A))]

Front: What is Bayesian updating?
Back:

• The process of updating beliefs about a state of the world in accordance with Bayes’ Rule.

Front: How does Bayesian updating work for a continuous state of the world?
Back:

• The world state is modeled as θ ~ N(m, σ²), where m is the mean and σ² is the variance.

• Precision ρ is defined as 1 / σ².

• A learner updates their belief after observing a noisy signal s about θ.

• Formula for updated belief: m' = βs + (1 - β)m, where β = ρₑ / (ρ + ρₑ).

Front: What is the Bayesian brain hypothesis?
Back:

• The hypothesis states that the brain receives noisy signals about reality and decodes them optimally by integrating prior knowledge.

• Estimation biases should change across contexts.

Front: What is the central tendency of judgment?
Back:

• Judgments gravitate towards the mean magnitude: lower values are overestimated, and higher values are underestimated.

• Explanation: noisy encoding + optimal decoding.

• The brain’s best guess of a stimulus magnitude is E[θ̂] = (1 - β)m + βθ.

• Model predicts underestimation for large θ and overestimation for small θ.

• Threshold: θ = m.

Front: What is the fallacy about independence?
Back:

• Assuming two things are independent when they are not.

• Example: Investing in stocks and bonds for diversification – perceived as independent but actually dependent.

• Example: The probability of getting heads after five tails – perceived as dependent but actually independent.

Front: What is the gambler's fallacy?
Back:

• The mistaken belief that a system has memory when it does not.

• Example: Believing a slot machine is "due" for a win.

Front: What is the hot hand fallacy?
Back:

• The belief that if someone is successful in a task now, they will continue to be successful in the future.

• Example: "I scored three basketball shots, I'll score the fourth."

Front: What is the representativeness heuristic?
Back:

• The probability of an outcome is judged by how representative it is of the process.

• Example: People incorrectly estimate the probability of coin sequences like HHHHHH < HHHHHP < HPPHHP.

• Leads to predictable biases and mistakes.

Front: What is the law of small numbers?
Back:

• The tendency to overestimate how well small samples represent a larger population.

• Example: "A sample of 4 is enough to prove the coin is biased."

Front: What is the conjunction fallacy?
Back:

• Overestimating the probability of a conjunction of events.

• Example: The Linda problem (banker vs. banker-activist).

• Example: Boeing has 6 million parts, each failing with a probability of 0.00001%. The probability that none fail is 0.25%, but people estimate it higher.

Front: What is the planning fallacy?
Back:

• Over-optimism in estimating the time required to complete a task.

• Example: PhDs take longer than planned due to optimistic assumptions.

Front: What is the disjunction fallacy?
Back:

• Underestimating the probability of at least one event occurring.

• Example: The probability of rolling at least one six with two dice.

• Related to the birthday problem.

Front: What is base-rate neglect?
Back:

• Ignoring the base rate of an event when considering conditional probabilities.

• Example: Cancer screening probabilities.

Front: What is confirmation bias?
Back:

• Tendency to interpret information in a way that supports existing beliefs.

Front: What is the availability heuristic?
Back:

• Assessing the probability of an event based on how easily examples come to mind.

• Explains:

  • Repeated dangerous behavior.

  • Why storytelling is effective.

Front: What is Sen’s α condition or Chernoff’s condition?
Back:

• If a decision-maker appears to prefer x to y in one menu, this preference should not be reversed in another menu.

• The choice is rationalizable if and only if it satisfies this property.

Front: What is the opportunity cost fallacy?
Back:

• Explicit cost: what you pay.

• Opportunity cost: the value to be forgone.

• People often disregard opportunity cost in decision-making.

Front: What is the sunk cost fallacy?
Back:

• The tendency to continue investing in something because of prior investment, even when it is not beneficial.

Front: What is the endowment effect?
Back:

• People tend to value what they already own more than what they don’t.

• A special case of status quo bias.

Front: What is loss aversion?
Back:

• Losses hurt more than equivalent gains feel good.

• Value function: Suppose the initial endowment is (m, w): V(m - m, w - w) = v(m - m*) + v(w - w*)**

• Loss aversion factor λ > 1, typically around 2.

Front: What is anchoring?
Back:

• In the absence of solid data, people use arbitrary reference points as anchors.

• Anchoring heuristic: If someone with no opinion hears an action suggested, they are likely to adopt it.

Front: What is the compromise effect?
Back:

• People tend to choose an option that is a compromise between two extremes.

Front: What is the decoy and asymmetric dominance effect?
Back:

• A tendency to choose an option that is strictly better than another, making it more attractive.

Front: What is the independence of irrelevant alternatives?
Back:

• If C is chosen from a set, it should still be chosen if a new option is introduced that does not affect its relative ranking.

• Introduction of a new point should not affect judgment.

Front: What is normalization in decision-making?
Back:

• The brain re-expresses input values relative to other inputs.

• Formula: a / (a + b).

Front: What is pairwise normalization in decision-making?
Back:

• Given a choice set X, an option x with {x₁, ..., xₙ} unnormalized values is evaluated as: V(x; X) = ∑(xₙ / (xₙ + yₙ))

• A decision-maker prefers x over y if V(x; X) > V(y; X).

Front: How does the brain apply normalization in decision-making?
Back:

• The brain compares input values relative to alternatives.

• Helps explain context-dependent choices.