Page 163 - Contributed Paper Session (CPS) - Volume 4
P. 163

CPS2165 Carla Susete Gonçalves Francisco et al.



                            Reaction times: A new approach using new
                                  advances in distribution theory
                                     ,1
                                                             2
             Carla Susete G. Francisco , José António S. Macias , Filipa Neiva C. Ribeiro 1
                                  1 Catholic University of Portugal, Lisbon, Portugal
                                      2 University of Coruna, Coruna, Spain
             Abstract:
             Reaction times (RTs) have been an important measure in the investigation
             of the cognitive process. Both psychologists and neurologists have been
             engaged in the development of studies on the evaluation of models to
             describe  the  RTs  processes.  A  very  important  issue  is  the  study  of
             probability  distributions  that  describe  the  RTs,  which  more  general
             probability  distribution  families  can  hardly  explain.  Among  the  several
             existing models, the most usual for this type of study is the LATER (Linear
             Approach to Threshold with Ergodic Rate) of Carpenter (1981). Nakahara
             et al. (2006) developed an extension of the LATER model, which they call
             the extended LATER model (ELATER), to account for trial-by-trial variability
             of both pre- and post-processes together. Moscoso (2008) develops a new
             theory in which TRs are directly proportional to the difficulty of the task
             and  inversely  proportional  to  the  rate  at  which  information  becomes
             available  to  solve.  The  possibility  of  fluctuations  might  be  taken  into
             account, in both the information gain rate (r) and in the resting level to the
             distance  (∆)  of  threshold  and  then  TRs  will  follow  a  distribution
             corresponding to the ratio between two normally distributed variables. The
             LATER  model  considers  that  the  values  of  (r)  and  (∆)  are  statistically
             independent of each other, although new pieces of evidence considered
             aren´t  correct,  once  several  experiments  produced  the  necessity  for
             introducing  an  additional  parameter  into  the  model  representing  the
             correlation between (r) and (∆). This project aims to develop a new model
             from  the  Fieller  distribution  for  the  ratio  of  two  normally  distributed
             variables. Finally, another issue that could cast doubts on the plausibility
             of  LATER  as  a  model  of  activity  accumulation  of  the  neurons  is  the
             constrained linear trajectory of the accumulation of evidence of validity. A
             reformulation  of  the  LATER  in  terms  of  the  diffusion  process,  named
             LATER-d was analyzed.

             Keywords
             Drift-diffusion  model;  Ergodic  Rate;  LATER;  Reaction  time;  Rise-to-
             threshold model.




                                                               152 | I S I   W S C   2 0 1 9
   158   159   160   161   162   163   164   165   166   167   168