The following images were a poster at the 1997 annual meeting of the Society for the Quantitative Analyses of Behavior.  The communication medium is intended for situations where the viewer is interactively stepped through the frames by the presenter.  The following is a one-sided version of the poster presentation.

Basic Idea

Most people are familiar with the characterization of the output of an audio amplifier (e.g., sound from a speaker) as the input signal (e.g., a microphone) plus amplification and distortion.




The same concept that was used to characterize the output of an amplifier can be applied to the characterization of the output of a living organism.  The behavior of an organism can be seen as the input plus the characteristics of that organism.  Alternatively stated, the distribution of the operant over time is the distribution of the reinforcers over time plus the characteristics of the organism.


Analytical and Numerical Methods

The productive use of that concept (i.e., that behavior is the input plus the characteristics of the organism) is mathematically complex.  The complexity can be dramatically simplified by carrying out the numerical operations necessary to predict behavior in a different domain.  Most people are familiar with a similar domain conversion to simplify a problem.


The multiplication of two numbers can be simplified to the addition of the logs of those two numbers.  The result (in the example "log C") is simply converted back to the original domain by taking the antilog.  This is the principle underlying "slide rulers" which were used by engineers before hand-held calculators.



 In the following example the transfer function for an organism is determined.  It tells us how the input was changed into the output by "passing through" the organism.  It is the characteristics of the organism.  In the same way that the output of an amplifier for any input can be predicted if the transfer function of the amplifier is known, in principle then, the output of an organism is known for any possible input if the transfer function for that particular organism is known.


The steps are as follows.  [left side] (blue upper left) the distribution of reinforcers provided to the organism (e.g., a 20-element Fleshler-Hoffman series with a mean of 60 seconds) is translated into the frequency domain (purple middle left).  Note that the conversion produces a set of values for both amplitude at each frequency and phase.  The output behavior of the organism which occurs as the result of that VI schedule input (orange upper right) is translated into the frequency domain (amber middle right).

 The frequency domain output (amber middle right) is divided by the frequency domain input (purple middle right).  The result is the frequency domain transfer function (green bottom center).
 Subsequently, [right side] any input (blue upper left) can be converted to the frequency domain (purple middle) multiplied by the previously obtained transfer function.  This results in a predicted output in the frequency domain (red middle right). This prediction can then be converted to the time domain (red upper right) and compared to the actual behavior (orange upper right).


Prior Research

We have already exposed birds to the necessary procedures, worked out the various procedural and numerical methods; and compared the obtained predictions to those obtained with actual organisms under novel test conditions.

The following panel presents a review of the rationale for the procedure, the actual experimental procedure, and the obtained results.


The first task of obtaining a useful transfer function for an audio amplifier or an organism is to expose the system to all possible frequencies.  In that way, a prediction can be made for the subsequent exposure to any possible combination of any possible input frequencies.  Obviously a brute force presentation of each frequency would be very time consuming.  The trick is to take advantage of the fact that a simple square wave or a "step function" actually contains all possible frequencies (proven by Fourierís theorem).  So if an organism is exposed to a VI 20" schedule for 200 seconds followed by extinction for 800 seconds, for example, then that is the same as presenting to that organism a whole series of different input frequencies individually.  This is in fact precisely what we did.

The procedure is illustrated in the middle of the figure.

It was shown that reasonable, zero-free parameter predictions could be made based on the use of a transfer function obtained with this procedure.  It is important to reiterate that no scaling adjustments are used with this method.  It is truly a zero free parameter prediction.   While it was shown that the conceptual machinery would work, the predictions were not perfect.  Our current research in this area is intended to further reduce the error of prediction caused by noise by developing better experimental procedures and better analytical and numerical methods.


Present Research: Determination of Transfer Function


The top portion of this frame illustrates a major source of artifactual noise in the determination of a transfer function with a step transition.  Division at each of the multiples of the reciprocals of the input step transition pulse duration is impossible because they are zero.  The loss of data at these points results in a noisy transfer function.

There is a solution to this problem.  It is illustrated in the "solution key" section of the figure.  If the zeros occur at reciprocals of the pulse duration, then different pulse durations have zeros at different points.  Exposing an organism to a two-pulse procedure would result in the cancellation of the zeros in the frequency domain version of the input because valid data is provided by one pulse duration or the other at every point across the spectrum.  This is precisely what we did.




The solution was to implement a procedure which provides 200 seconds of VI 20-sec followed by 231 seconds of extinction followed by 141 seconds of VI 20-sec followed by 1428 seconds of extinction.  The trial start is designated by a 10 second blackout.  The blackout is provided as a temporal anchor, because there is no explicit signal designating the VI periods from the extinction periods (i.e., it is a mixed schedule).

Present Research Prediction of Behavior

One of the great benefits of this analytical approach to behavior analysis is that it gives us a way to understand one of the most fundamental questions in the analysis of behavior.  A familiar example of this issue is  given by bandpass filters.

A broad-band filter passes all the details (both sharp edges and slow trends) of the input signal.  An audio amplifier is better if it passes all of these details.  A low pass filter on the other hand, would blur all the fine details and give only the broad averages.  A very inexpensive or low fidelity amplifier is said to have a poor "frequency response" because of its poor band pass characteristics.




 An important aspect of the behavior of organisms is the degree to which it follows all the small transitions in the contingencies of reinforcement or the degree to which is smoothes over the fine details.  The most obvious example of smoothing over changes in the input is provided by the behavior under a VI schedule.  Even though there are periods of reinforcers far apart as well as close together,  the organism has a steady output.  This is illustrated in the bottom section of the above illustration. This difference in responsiveness to the details in the contingencies of reinforcement has been labeled "molecular" (reacts to small changes) and "molar" (does not react to small changes).

This difference can be more usefully seen as the filter characteristics of the organism (e.g., broad-band versus low-pass).  The importance of this relabeling is that it connects us to an enormous body of conceptual and analytical tools used to great benefit by "older" sciences.

The great power of a transfer function analysis is that it will predict a zero free parameter output for any possible input.  It will specify the precise filter characteristics for particular organisms in particular situations given even complex input signals.




In our case, we chose to use a test procedure to evaluate the accuracy of the transfer function which would allow us to produce precisely a 20% sag in behavior and to produce precisely an 80% elevation in behavior.  This was done with a three pulse procedure.  The separation of the first two VI pulses was adjusted such that the transfer function predicted a 20% sag in rate for the test bird.  The width of the third VI pulse was adjusted such that the transfer function predicted an 80% rise in rate for that bird.  No currently available model can make a similar, principled, quantitative prediction about the degree to which an organism is "molecular" or "molar."

Supplemental Numerical Analytical Treatments

We were able to uncover and remove several additional procedural and analytical sources of artifactual noise.  One was caused by our choice of procedures.  A VI pulse was in effect immediately following the blackout.  This resulted in an instantaneous rate change from zero responses per second in the blackout to a high rate during the VI schedule.  This large amplitude instantaneous change in output caused high frequency noise in the transfer function.


As a kluge for one bird (because there was no time to collect data for a new transfer function and include it in the scheduled poster session) we simply copied the gradual rate increase from the second slope to the first slope.  This eliminated that source of noise.

Our current solution is to use a procedure which does not result in an instantaneous, large magnitude rate change.  The first VI pulse is offset from the trial start so that behavior will not instantaneously change to a high rate.




As a second noise reduction technique we converted the transfer function (frequency domain) to an impulse response function (time domain) filtered the high frequency noise (which is easy in the time domain), then translated the result back to the frequency domain.  This filtered transfer function was then used to do two things:  (1) generate the durations of the VI schedules and the extinctions which would produce exactly a 20% sag and exactly an 80% elevation in response rate for that particular bird, and (2) generate an exact predicted output for that bird for each point in time for that three-pulse procedure.

Our current solution is to no longer allocate so much of the spectrum to behaviorally impossible frequencies.  We are no longer sampling at an extremely high rate which "detects" substantial amounts of high frequency noise and then subsequently removes that noise with complex numerical methods.  By analogy, rather than measuring the time of pecks to the picosecond (which results in substantial artifactual variability in the peck items) and then smoothing down to milliseconds, we are measuring peck times to the nearest millisecond.  For example, it is simply not meaningful to emphasize picosecond variability if pecks were to occur exactly 38 milliseconds apart.



Results: Prediction and Obtained Behavior

The following is the obtained output for the test bird on the three-pulse procedure and the behavior predicted by the transfer function.

Obtained:

Predicted:

The important characteristics to note are:

  1. the amount of sag between the first two pulses
  2. the amount of elevation in responding to the third pulse
Additional points to note are the fine grain of the behavior and the prediction, as well as the "ringing" predicted by the transfer function and the single "ring" in the obtained behavior.

While the transfer function predicts extinction "below zero," our procedure was not capable of substantiating its existence.



Finally, as a terminological issue several types of behavior dynamics are illustrated.
 

The changing behavior to a synchronous, repetitive task (i.e., which can be averaged over many instances) such as the behavior change under an FI schedule, or our two-pulse procedure, are labeled synchronous dynamics because they are synchronized to events in nature.  This is much like an oscilloscope sweep which can be synchronized to some event such that the waveform is repetitively presented in the "same place" on the screen.  In our previous paper, this was labeled repeated measures local average.

Transient dynamics are well known as "behavioral adaptation" or "the learning curve."  These are not synchronous in that the function is a one time change.

Finally, the most poorly understood behavior dynamics are asynchronous dynamics.  These changes are sometimes referred to as random unsystematic fluctuations in what is otherwise stable behavior.  The grand challenge is to change as much of this residual variance to "variance accounted for" as is possible.

A detailed analysis of synchronous dynamics was presented by Palya (1992) in Dynamics and the Fine Structure of Behavior. The animated figure below illustrates the asynchronous dynamics in our two-pulse procedure for bird 568 from session 3 to session 192.  Each of the 38 frames is the average of 5 sessions.


Results: Asynchronous Dynamics

  



SEBAC Psychology Server

Send comments/criticisms/speculations to
Date Last Reviewed : January 10, 2002