Computational Cognitive Modeling: The Drift Diffusion Model
Sequential Sampling Models
The Wiener Diffusion Process
The Drift Diffusion Model (DDM) separates two-choice decision processes into distinct cognitive components by modeling reaction time and accuracy simultaneously. Evidence accumulates over time according to a stochastic Wiener process: where is the drift rate (signal) and represents intra-trial Gaussian noise.
Drift Diffusion vs. Reinforcement Learning
Drift Diffusion Model (DDM)
- Models rapid, single-trial perceptual decisions
- Focuses on reaction time distributions and accuracy
- Parameters represent evidence accumulation speed and thresholds
Reinforcement Learning (RL)
- Models learning across multiple trials
- Focuses on choice probabilities and value updating
- Parameters represent learning rate and exploration/exploitation
Parameter Estimation
Estimating DDM parameters using Maximum Likelihood can be unstable with small trial counts. Modern approaches use Hierarchical Bayesian estimation (e.g., HDDM in Python) to constrain individual parameters using group-level priors.
Which parameter in the Drift Diffusion Model primarily reflects the speed of information processing or the quality of the stimulus?
The drift rate () quantifies the average rate of evidence accumulation. Higher drift rates indicate faster and more accurate information processing.
Correct answer: Drift rate (v)
Which DDM parameter represents the distance between the two decision thresholds, reflecting response caution?
Boundary separation () dictates how much evidence is needed before a decision is made. A wider boundary leads to slower, more cautious, and more accurate decisions.
Correct answer: Boundary separation
DDM Parameters and Cognitive Interpretations
| Parameter | Symbol | Cognitive Interpretation |
|---|---|---|
| Drift Rate | v | Speed of evidence accumulation / task difficulty |
| Boundary Separation | a | Speed-accuracy tradeoff / response caution |
| Non-decision Time | t or Ter | Motor execution and sensory encoding time |
| Starting Point | z | A priori bias toward one response alternative |