An update on analytical estimates of limited sampling biases in different information measures

We revisit the 1996 Network: Computation in Neural Systems article "Analytical estimates of limited sampling biases in different information measures". We discuss why sampling bias was a major obstacle to the empirical study of neural coding. We review how our analytical approach clarified the limitations of earlier corrections and provided robust bias-corrected information estimators that helped open the way to study spike timing and information in neural populations. We finally consider its continuing relevance to modern neural coding studies.