Re: Pavlovian Conditioning and Cognitive Lockup

From: Harnad, Stevan (harnad@soton.ac.uk)
Date: Fri Feb 24 1995 - 14:16:42 GMT


> From: "Young, Mark" <MYOUNG92@psy.soton.ac.uk>
> Date: Fri, 24 Feb 1995 14:10:12 GMT
>
> The Ader & Cohen material on Pavlovian conditioning can apply to a
> lot of areas - the diabetics, the hungry and the alcoholic. However,
> it works the other way for the depressed. The biological account
> posits that low levels of serotonin is a cause for depression. But
> antidepressants, which raise serotonin levels very quickly, do no
> "work" for 10-14 days. Also, the drugs stop boosting
> neurotransmitter levels after a few days, when serotonin levels drop
> back to normal. Yet mood lifts a week or so later.

The understanding of the mechanisms of mood is minimal. Nothing like a
coherent cause/effect explanation is in sight yet, molecule-by-mood.
Yet we know that certain antidepressent drugs (lately the serotonergic
ones) work. The ill-understood 10-14 day lag is only the tip of the
iceberg when it comes to what is NOT understood about all this, and it
presents itself as such a prominent mystery probably only because the
REST of our simple cause/effect ideas about what's going on here are
wrong or incomplete too.

> The fact that, in the Ader & Cohen expt, the effects of saccharin
> decay after 4 or 5 trials without the UCS can be explained by normal
> extinction, as identified by Pavlov. Presumably, similar phenomena
> such as spontaneous recovery would also occur. The reason this would
> not work as an alternative to therapy is that the CR fades over time,
> such that it may not be effective at a time when CS treatment becomes
> an option.

Correct.

> The use of counterconditioning and desensitisation may be effective,
> although I'm curious: does the exploitation of extinction itself
> extinguish (i.e. will there be a spontaneous recovery of the phobia)?

Here's some material that may help:

    Hugdahl, Kenneth; Fredrikson, Mats; Ohman, Arne. Preparedness and
    arousability as determinants of electrodermal conditioning.
    Behaviour Research & Therapy, 1977, v15 (n4):345-353.

    ABSTRACT: Arousability, as defined through spontaneous
    electrodermal responses, has been empirically linked to anxiety,
    phobic symptoms, and outcome of systematic desensitization.
    Previous data indicate that preparedness, as defined through
    potentially phobic vs fear-irrelevant or neutral conditioned
    stimuli, is an important determinant of electrodermal
    conditioning. The present experiment compared 4 groups of 16
    undergraduates selected to be high or low in spontaneous responding
    during differential conditioning to potentially phobic or neutral
    stimuli. It was found that the effects of these 2 factors were
    essentially additive, i.e., conditioning and resistance to
    extinction were better for phobic stimuli and for high-arousal
    groups. The high-aroused group with phobic stimuli showed diffuse
    responding during acquisition, not differentiating between
    reinforced and unreinforced cues. However, it was the only group
    that failed to extinguish during 20 trials, which indicates that
    high arousal gives superior resistance to extinction particularly
    for phobic stimuli.

To retrieve the following article, use this URL:
ftp://cogsci.ecs.soton.ac.uk/pub/harnad/BBS/bbs.davey

    PREPAREDNESS AND PHOBIAS: SPECIFIC EVOLVED ASSOCIATIONS OR A
    GENERALIZED EXPECTANCY BIAS?

                Graham Davey
                Psychology Division,
                Department of Social Science,
                The City University,
                Northampton Square,
                LONDON EC1V 0HB U.K.
                J.A.Fildes@city.ac.uk
 

    ABSTRACT Most phobias are focussed on a small number of
    fear-inducing stimuli (e.g. snakes, spiders). A review of the
    evidence supporting biological and cognitive explanations of this
    uneven distribution of phobias suggests that the readiness with
    which such stimuli become associated with aversive outcomes arises
    from biases in the processing of information about threatening
    stimuli rather than from phylogenetically based associative
    predispositions or "biological preparedness." This cognitive bias,
    consisting of a heightened expectation of aversive outcomes
    following fear-relevant stimuli, generates and maintains robust
    learned associations between them. Some of the features of such
    stimuli which determine this expectancy bias are estimates of how
    dangerous they are, the semiotic similarity between them and their
    aversive outcomes, and the degree of prior fear they elicit.
    Ontogenetic and cultural factors influence these features of
    fear-relevant stimuli and are hence important in determining
    expectancy bias. The available evidence does not exclude the
    possibility that both expectancy biases and specific evolved
    predispositions co-exist, but the former can explain a number of
    important findings that the latter cannot.

    KEYWORDS: Preparedness, phobias, biological preparedness, selective
    associations, information processing biases, classical
    conditioning, covariation assessment, cross-cultural studies.

> This still does not explain the anomalous depression results
> explained above - any ideas?

What is not explained in the depression results is a lot more than the
time lag; indeed, the fact that the time lag appears anomalous may just
be because of our simplistic and incomplete understanding of the full
causal mechanism involved.

> Desperately trying to relate this to circles now. Perhaps by
> relating it to established research, we avoid falling into this
> circle, which is essentially just a version of cognitive lockup (that
> Sherlock Holmes thing again - fitting facts to theory rather than the
> other way round). This is not a new idea - anyone (apart from
> Psych/Law students) heard of the cognititive interview?

Hermeneutics comes into the work on Pavlovian Conditioning of the Immune
Response when one extrapolates the trivial laboratory results on mice to
imagined real effects in the clinic (passing over the sticky problem of
extinction and effect size).

Some food for thought: so called "theory-saving" (the far-fetched
re-interpretation of seemingly negative results so as to support your
preferred theory) is a hermeneutic effect, even though it can happen
with an empirical theory too! In fact, this is the soft underbelly of
empiricism: Your THEORY may be empirical and falsifiable, but to the
extent that there is room for interpretation in both the theory and the
data (and how/why they fit), you could still be spinning around in
hermeneutic circles (not unlike Ptolemaic Epicycles, for those of you
who know the famous example of the earth-centered theory whose
proponents tried to save it against the Copernican Revolution by
inventing more and more far-fetched "epicycles" to explain the finding
that planets moved in eliptical paths rather than the circular ones
Ptolemy had posited).

Cure: Big, reliable results, supported by the literature, both
experimental and theoretical. This is not to say that revolutionary
results cannot and will not occasionally appear unexpectedly out of
left field, but you cannot play it as if that were the usual effect,
rather than the rare outlier. Otherwise you risk, like Polonius, going
completely hermeneutic, interpreting the amplified noise between the
radio stations as if it were whispering your name. Rather keep
looking for the reliable signal of the station.

Remember the old joke about conditional probability: Mrs. Jones is
WRONG to think that the (alleged) fact that Einstein was a slow learner
is good news for her own little late-bloomer, Jonas: If sometimes X
happens DESPITE Y, it does not follow that if Y is the case, X is any
more likely.

Chrs, Stevan



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:15 GMT