clockdata_run-class: dataset object for run-level data (multiple trials)

In PennStateDEPENdLab/fitclock: An R package to fit RT and Reward behavioral data from Michael Frank clock task.

Description

Note that the workspace used during the fitting process resides at the level of the clockdata_run object since prediction equation is principally intended to fit multiple trials in a given run. Optimal parameters across runs or subjects and runs are derived by summing SSEs for each run, essentially finding a single set of parameter values that fit all runs/subjects reasonably well.

Fields

w:

shared workspace environment used during fit.

run_number:

number of this run in a multi-run sequence. Important when some values (e.g., V) carry across runs.

SSE:

sum of squared errors for this run: predicted versus observed RTs

RTobs:

vector of observed RTs

Reward:

vector of obtained rewards.

avg_RT:

average reaction time for this run (maybe overridden by global RT)

global_trial_number:

vector of trial numbers in the overall experiment (1..runs x trials/run)

rew_function:

string denoting run reward contingency (e.g., "IEV")

run_condition:

string denoting some other aspect of run (e.g., emotion, such as "happy")

by_lookup:

at fit-time, clock_model copies in a named character vector that is the union of all relevant fields for run definition (usually rew_function + run_condition)

orig_data_frame:

optional data.frame from original experiment run containing full saved data (in case there are additional variables of interest)

Methods

initialize_workspace(prior_w):

sets up shared environment for fitting, w, based on prior_w, workspace of prior run.

plot_RTs():

plot observed RTs (partially implemented)

PennStateDEPENdLab/fitclock documentation built on May 8, 2019, 1:29 a.m.