Pupillometry deconvolution methods
This is a methods paper on deconvolution of pupillometry data for auditory experiments.
McCloy D, Larson E, Lau B, & Lee AKC (2016). Temporal alignment of pupillary response with stimulus events via deconvolution. The Journal of the Acoustical Society of America, 139(3), EL57–EL62. doi: 10.1121/1.4943787
Raw data is cleaned and aggregated with
analyze-data.py (for the
pupil impulse response experiment) and
analyze-voc-data.py (for the vocoded
letters experiment). These should create the summary data objects needed to
create the figures (
avg_data.npz for Figure 1;
voc_data_wierda.npz for Figure 3).
Once those are in place, the makefile for the article accepts directives
(for prepress version formatted similar to the final formatting in JASA-EL),
sub (for the JASA-EL submittable version: with double spacing, line numbers,
list of figures, etc), or
web (for a prepub manuscript with my preferred
formatting for posting on the web).
Figure generation is not 100% automated for the
make sub directive, because it
requires opening the auto-generated PDF figures in Adobe Illustrator and saving
them as EPS files in order to avoid rasterization of the semi-transparent
regions of the plot (only strictly necessary for Figure 3).