Spectro-temporal receptive field


The neuron's acoustic response


Auditory neurons may alter their firing response to changes of stimuli in space, time and frequency. This is characterized with the receptive field of a neuron. Here we will derive the spectral-temporal receptive field (STRF) of an primary auditory cortex neuron, which describes the response of that neuron to changes in sound frequency across time. One can derive the STRF in several ways, but we will implement a method that is also used in a study on response patterns of inferior colliculus neurons, determining the STRF from the neuron's responses to rippled sounds.


The data for this tutorial can be found here. The PANDA toolbox is not required, but can be useful. The data for this tutorial is derived from an experiment in which the awake subject listened to ripples, while a single neuron from primary auditory cortex was recorded with an electrode. Spikes in the data have already been detected and sorted for your convenience.


The ripple stimuli consisted of a broadband complex of tones, with the ripple envelopes being sinusoidally modulated in the spectrotemporal domain. Figure 1 shows an example amplitude-modulated noise, with an envelope modulation across time, defined by the parameter ripple velocity \( \omega = 4 \) (cycles/sec). The rippled sound starts after 500 ms of static noise (this part of the sound is not essential for determing an STRF; it is just part of the experimental paradigm employed). The parameter that defines the modulation across frequency is termed ripple density, \( \Omega \) (cycles/octave). A set of 55 different ripples was presented, and consisted of 5 different velocities (8:8:40 Hz) and 11 different ripple densities (–2.0:0.4:+2.0 cyc/oct).



We first need to read the data into MatLab.

cd('C:\DATA\'); % go to data-directory
load('examplecell'); % load data

which will yield a structure in Matlab's workspace called spike, with fields like spiketime, stimparams, and stimvalues:

>> whos
Name       Size              Bytes  Class     Attributes
  spike      1x274            508736  struct              
>> spike
1x274 struct array with fields:

Spectro-temporal receptive field analysis

Next, we will make a raster plot and determine the peri-stimulus time histogram, to see whether we can observe a general change in the spiking activity of the neuron to this acoustic stimulation.