ABSTRACT
Wedescribetheconceptualbackgroundandpracticalimplementationsof sometechniquesfortheanalysisofsymbolsequencesandsymbolictime series.Weemphasizetheirassociatedsoftwarerealization,theWinGramm suiteofprograms,thatincludesprogramsforthecalculationofconditional entropies,context-freegrammaticalcomplexity,andalgorithmicdistance andredundancy,aswellasforthegenerationofsurrogatesthatpreserve symbolpairsandtriplets.Wedemonstratetheusefulnessoftheseprograms bymeansoftwoillustrativeexamples,takenfromcomputationalneuroscience.Inthe‡rstone,weobtainevidenceoftheMarkoviancharacterofthe corticalinter-spikeintervals(ISIs)oftheratbeforepenicillintreatment,and itsdisappearanceafterward.Inthesecondone,weextendpreviousinvestigationsaboutneuralspiketrainsgeneratedbytheisolatedneuron of the
CONTENTS
Materials and Methods....................................................................................... 475 Entropy-Like Measures of Sequence Structure .......................................... 475 Context-Free Grammatical Complexity ......................................................477 Algorithmic Redundancy .............................................................................. 478 Surrogate Sequences ...................................................................................... 479 Algorithmic Distance ..................................................................................... 479
Construction of Symbolic Partitions ................................................................480 Applications ......................................................................................................... 481 Conclusions ..........................................................................................................488 Appendix 20.A ..................................................................................................... 491 Acknowledgments .............................................................................................. 492 References ............................................................................................................. 492
slowly adapting stretch receptor organ (SAO), in order to classify sequences of different lengths of known neural behaviors. We include new spike trains, digitized employing the optimal partition procedure described by Steuer, Molgedey, Ebeling, and Jiménez-Montaño (2001).
