Andrew Heathcote, working at the University of Tasmania and later the University of Newcastle (Australia), has been one of the most productive contributors to the mathematical modeling of response times and decision processes. His co-development of the Linear Ballistic Accumulator (LBA) model with Scott Brown provided an elegantly simple yet powerful account of multi-alternative choice, and his work on Bayesian estimation methods has made evidence accumulation models accessible to a broad research community.
The Linear Ballistic Accumulator
Starting point: k_i ~ Uniform(0, A)
Drift rate: d_i ~ Normal(v_i, s)
Decision time: T_i = (b - k_i) / d_i
Response: choose i with smallest T_i
The LBA's key innovation is that accumulation within a trial is deterministic and linear -- all variability comes from between-trial variation in starting points and drift rates. This makes the model mathematically tractable (closed-form likelihood functions), computationally fast to fit, and naturally extensible to any number of alternatives. Despite its simplicity, the LBA produces the full range of benchmark phenomena: right-skewed RT distributions, speed-accuracy tradeoffs, and correct responses being faster than errors on average.
Heathcote has been a leader in developing hierarchical Bayesian methods for fitting evidence accumulation models, including both the LBA and diffusion models. His Bayesian approach estimates individual-level parameters while constraining them with group-level distributions, providing principled handling of individual differences and enabling the application of accumulation models to clinical and developmental populations.
Response Time Distribution Analysis
Heathcote has contributed extensively to methods for analyzing RT distributions. His work on fitting the ex-Gaussian distribution, comparing distributional models, and developing robust estimation methods has made distributional analysis practical for experimentalists. His QMPE (Quantile Maximum Probability Estimator) addresses estimation challenges with right-skewed, contaminated RT data.
Legacy and Impact
Heathcote has been instrumental in making formal cognitive modeling practical and accessible. The LBA demonstrated that mathematical elegance and empirical adequacy can coexist -- simple enough to teach in graduate courses yet powerful enough for serious research. His development of fitting tools, hierarchical estimation methods, and software packages has lowered the barrier to entry for evidence accumulation modeling, helping these models become the standard framework for analyzing decision processes in cognitive psychology and neuroscience.