Study Notes on Classical Conditioning

Definition: Classical Conditioning (CC) is a learning process where a neutral stimulus (NS) comes to evoke a response after being paired with a stimulus that naturally evokes that response.
Key Components:
1. Unconditioned Stimulus (US): A stimulus that naturally and automatically triggers a response without prior learning (e.g., food).
2. Unconditioned Response (UR): The unlearned, naturally occurring response to the US (e.g., salivation).
3. Conditioned Stimulus (CS): An originally neutral stimulus that, after association with a US, comes to trigger a conditioned response.
4. Conditioned Response (CR): The learned response to a previously neutral (but now conditioned) stimulus.
Example (Food Aversion):
- Initial state: Crab (NS) + Stomach flu (US) -> Nausea (UR).
- Result: Crab (CS) -> Nausea (CR). This is often called the Garcia Effect, demonstrating that some associations (like taste-illness) are formed more easily due to biological preparedness.

Procedural Overview: In humans, a CS (like a $1000 \text{ Hz}$ tone) is paired with a US (a brief air puff to the cornea). Over trials, the subject begins to blink (CR) just before the air puff (US) occurs.
Temporal Dynamics:
- Delay Conditioning: The CS overlaps with the US and lasts until the US is presented. This is generally the most effective method.
- Trace Conditioning: A brief interval (the "trace") exists between the end of the CS and the onset of the US. This requires a memory trace and often involves higher-order brain structures like the hippocampus.
Optimal Timing: Conditioning is most successful when the CS-US interval ranges between $250$ and $700 \text{ msec}$ . If the interval is too short ( $< 50 \text{ msec}$ ) or too long (several seconds), associative learning drops significantly.

Habituation: A decrease in response strength following repeated exposure to a non-threatening stimulus. This allows organisms to filter out irrelevant information.
Sensitization: An increase in response strength following exposure to a noxious or intense US (e.g., a loud noise or sharp air puff). This serves as a general arousal mechanism to heighten vigilance.

Invertebrate Models: In Aplysia californica, conditioning involves changes in the synaptic strength between sensory neurons and motor neurons, mediated by the release of neurotransmitters (e.g., serotonin) that increase calcium influx.
Vertebrate Models:
- Cerebellum: Crucial for eyeblink conditioning. The Interpositus Nucleus is the site where the association between the CS and US is stored.
- Amygdala: Essential for fear conditioning (pairing a tone with a shock).

S-R (Stimulus-Response): Assumes a direct link is formed between the CS and the UR pathway. The CS triggers the response directly without needing to activate the "representation" of the US.
S-S (Stimulus-Stimulus): Assumes the CS activates a representation of the US, which then triggers the response.
US Devaluation: To test between S-S and S-R, researchers condition an animal and then "devalue" the US (e.g., satiate the animal or pair the US with illness).
- If CR decreases, it suggests S-S learning (the animal expects the US but no longer wants it).
- If CR remains high, it suggests S-R learning (the response is reflexive and independent of the current value of the US).

Sensory Preconditioning:
1. Pair two neutral stimuli: Stimulus A and Stimulus B (no response yet).
2. Condition Stimulus B with a US to produce a CR.
3. Test Stimulus A; if it elicits a CR, learning occurred during the first phase.
Second-Order Conditioning:
1. Condition CS1 with a US.
2. Pair a new CS2 with CS1.
3. CS2 eventually elicits the CR despite never being paired with the US.

This model suggests that conditioning is about "surprise" or prediction error. If the US is exactly what was expected, no further learning occurs.
Mathematical Formula: $\Delta V<em>i = \alpha</em>i \beta (\lambda - \sum V)$
- $\Delta V_i$ : The change in associative strength of CS $i$ .
- $\alpha_i$ : Salience of the CS.
- $\beta$ : Learning rate associated with the US.
- $\lambda$ : The maximum associative strength that the US can support.
- $\sum V$ : The total associative strength of all stimuli present on a trial.
Blocking Effect: If a US is already fully predicted by CS1 ( $V = \lambda$ ), then pairing it with a new CS2 results in no learning for CS2 because the prediction error is zero ( $\lambda - \sum V = 0$ ).

CC is not just a laboratory phenomenon; it is an evolutionary adaptation. It allows organisms to predict vital events like food availability, predator presence, or mating opportunities, providing a "head start" in reacting to biologically significant stimuli.