Modulation patterns are known to carry critical predictive cues to signal detection in complex acoustic environments.  The current study investigated the persistence of masker modulation effects on post-modulation detection of probe signals.  Hickok et al. (2015) demonstrated that thresholds for a tone pulse in stationary noise follow a predictable periodic pattern when preceded by a 3-Hz amplitude modulated masker.  They found entrainment of detection patterns to the modulation envelope lasting for approximately 2 cycles after termination of modulation.  The current study extends these results to a wide range of modulation rates by mapping the temporal modulation transfer function for persistent modulatory effects.  We found significant entrainment to modulation rates of 2 and 3 Hz, a weaker effect at 5 Hz, and no entrainment at higher rates (8 to 32 Hz).  The effect seems critically dependent on attentional mechanisms, requiring temporal and level uncertainty of the probe signal.  Our findings suggest that the persistence of modulatory effects on signal detection is lowpass in nature and attention based.

Data were collected using Matlab software on a Sony Lenovo T400 computer in a steel-walled acoustically isolated chamber (Industrial Acoustics Company).  On each trial of a single-interval two-alternative forced-choice task, the subject was required to indicate (via a key press) whether a tonal signal was present during the unmodulated segment of a masking noise. Feedback was provided after each trial.  The data made available here is the raw data in a .mat file which includes results of every single trials of every run by every subject.  The files include the raw data without any postprocessing.   

The entire raw dataset for this paper is made available here.   Folders named “Fig 2” and “Fig 3” contain data used to generate figures 2 and 3 in the published manuscript.

The Excel file “Subject list” provides a list of participants that shows which subject (S1 to S9) participated in which condition of each experiment.

Each data file in each folder is in .mat format which can be accessed using Matlab software.  Each file contains an nx5 data format (n rows by 5 columns).  Each row represents one trial of the experiment for the condition identified by the file name.  For example, for Figure 2, the file named S1perdata2hz contains data from subject S1 in the periodicity (per) experiment at the 2-Hz noise modulation condition.  This particular file contains 3000 rows (trials) by 5 columns. 

The columns represent:

Column 1: Signal (tone pulse) onset delay in milliseconds relative to the offset of the noise modulation. The zero point is arbitrarily set at the expected dip of the first cycle (in the steady state part of the noise) after termination of noise modulation.   The smallest negative number (delay) represents the onset of the signal synchronized with the offset of the noise modulation. 

Column 2:  Scalar weight by which the signal (tone) was multiplied.  There were 5 levels which spanned a range of approximately 12 dB.  Data from some initial pilot runs at other test levels are also included in some data files, but not used in generating figures.

Column 3:  Designates whether a signal was present on that given trials (single interval forced choice). 1= signal present, 0= no signal present.

Column 4: Subject’s response (1 = yes there was a signal presented on this trial, 0 = no there wasn’t a signal presented on this trial).

Column 5: Was the subject’s response correct on this trial (i.e., whether columns 3 and 4 matched).  1= correct, 0= incorrect.