Manufacturing PTSD evidence with machine learning

What would you do if you were a scientist who had strong beliefs that weren’t borne out by experimental evidence?

Would you be honest with yourself about the roots of the beliefs? Would you attempt to discover why the beliefs were necessary for you, and what feelings were associated with the beliefs?

Instead of the above, the researchers of this 2017 New York human study reworked negative findings of two of the coauthors’ 2008 study until it fit their beliefs:

“The neuroendocrine response contributes to an accurate predictive signal of PTSD trajectory of response to trauma. Further, cortisol provides a stable predictive signal when measured in conjunction with other related neuroendocrine and clinical sources of information.

Further, this work provides a methodology that is relevant across psychiatry and other behavioral sciences that transcend the limitations of commonly utilized data analytic tools to match the complexity of the current state of theory in these fields.”

1. The limitations section included:

“It is important to note that ML [machine learning]-based network models are an inherently exploratory data analytic method, and as such might be seen as ‘hypotheses generating’. While such an approach is informative in situations where complex relationships cannot be proposed and tested a priori, such an approach also presents with inherent limitations as a high number of relationships are estimated simultaneously introducing a non-trivial probability of false discovery.”

2. Sex-specific impacts of childhood trauma summarized why cortisol isn’t a reliable biological measurement:

“Findings are dependent upon variance in extenuating factors, including but not limited to, different measurements of:

  • early adversity,
  • age of onset,
  • basal cortisol levels, as well as
  • trauma forms and subtypes, and
  • presence and severity of psychopathology symptomology.”

Although this study’s authors knew or should have known that review’s information, cortisol was the study’s foundation, and beliefs in its use as a biomarker were defended.

3. What will it take for childhood trauma research to change paradigms? described why self-reports of childhood trauma can NEVER provide direct evidence for trauma during the top three periods when humans are most sensitive to and affected by trauma:

The basic problem prohibiting the CTQ (Childhood Trauma Questionnaire) from discovering likely most of the subjects’ historical traumatic experiences that caused epigenetic changes is that these experiences predated the CTQ’s developmental starting point.

Self-reports were – at best – evidence of experiences after age three, distinct from the experience-dependent epigenetic changes since conception.”

Yet the researchers’ beliefs in the Trauma History Questionnaire’s capability to provide evidence for early childhood traumatic experiences allowed them to make such self-reports an important part of this study’s findings, for example:

“The reduced cortisol response in the ER [emergency room] was dependent on report of early childhood trauma exposure.” “Utilization of machine learning for prediction of post-traumatic stress: a re-examination of cortisol in the prediction and pathways to non-remitting PTSD”


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.