I Wanna Hold Your Hand (after 23 sessions of Emotionally Focused Therapy)

Can neuroscience illuminate the nature of human relationships? Or does it primarily serve as a prop to sell self-help books? The neurorelationship cottage industry touts the importance of brain research for understanding romance and commitment. But any knowledge of the brain is completely unnecessary for issuing take-home messages like tips on maintaining a successful marriage.

In an analogous fashion, we can ask whether successful psychotherapy depends on having detailed knowledge of the mechanisms of “neuroplasticity” (a vague and clichéd term). Obviously not (or else everyone's been doing it wrong). Of course the brain changes after 12 sessions of psychotherapy, just as it changes after watching 12 episodes of Dexter. The important question is whether knowing the pattern of neural changes (via fMRI) can inform how treatment is administered. Or whether pre-treatment neuroimaging can predict which therapy will be the most effective.

However, neuroimaging studies of psychotherapy that have absolutely no control conditions are of limited usefulness. We don't know what sort of changes would have happened over an equivalent amount of time with no intervention. More importantly, we don't know whether the specific therapy under consideration is better than another form of psychotherapy, or better than going bowling once a week.

Enter Love Sense: The Revolutionary New Science of Romantic Relationships, a new book by Dr. Sue Johnson, the clinical psychologist who developed Emotionally Focused Therapy (EFT).¹ The book is reviewed by Dr. Helen Fisher in the New York Times:

Love in the Time of Neuroscience

By HELEN FISHER  FEB. 7, 2014

In “The Devil’s Dictionary,” Ambrose Bierce defined love as “a temporary insanity curable by marriage.” Enter Sue Johnson, a clinical psychologist and couples therapist who says that relationships are a basic human need and that “a stable, loving relationship is the absolute cornerstone of human happiness and general well-being.” To repair ailing partnerships, she has developed a new approach in marriage counseling called Emotionally Focused Therapy, or EFT, which she introduces in her new book, “Love Sense.”
...

Johnson believes EFT can help couples break out of patterns, “interrupting and dismantling these destructive sequences and then actively constructing a more emotionally open and receptive way of interacting.” She aims to transform relationships “using the megawatt power of the wired-in longing for contact and care that defines our species,” and offers various exercises to restore trust.

Most interesting to me was Johnson’s brain-scanning study. Before EFT therapy, unhappily married women participating in the study reported considerable pain from an electric shock to the ankle as they held their husbands’ hands. After 20 sessions of EFT, however, these now more securely attached women judged their pain as only “uncomfortable” and their brain scans showed no alarm response. Secure attachment appears to change brain function and reduce pain.

Initial questions:

• Is there a “wired-in longing for contact and care that defines our species”? {my needy cat seems to long for contact and care}

• What's with that hand-holding ankle shock brain-scanning study? {did EFT really eliminate the “alarm response” in these women?}

 
Then Fisher continues:

But Johnson too often focuses on attachment to the exclusion of other “megawatt” brain systems. Remarkably, she lumps romantic love with attachment, saying “adult romantic love is an attachment bond, just like the one between mother and child.” In reality, romantic love is associated with a constellation of thoughts and motivations that are strikingly different from those of attachment. My research bears out that humankind evolved distinct but interrelated brain systems for mating and reproduction: the sex drive (to seek a range of partners); feelings of romantic love (to focus one’s mating energy on a single partner); and feelings of attachment (to drive our forebears to form a pair-bond to rear their young together). Each brain system is associated with different neurochemicals; each is a powerful drive that still plays a continuing role in partnership stability.

More questions:

• Are there distinct (but interrelated) brain systems for the sex drive, romantic love, and feelings of attachment? {I actually find this to be plausible}

• Is each brain system associated with different neurochemicals? {i.e. testosterone, dopamine, and oxytocin, respectively. I find this to be less plausible, or at least a bit simplistic.}

It's time to correct the misperceptions and overinterpretations that have arisen from this research!!

This is a job for...

Magneto

This looks like a job for @sarcastic_f! (I imagine him as some sort of BS-fighting super hero) http://t.co/z29q00G80u
— Adam J Calhoun (@neuroecology) February 8, 2014

Since there are a number of issues to tackle here – too many for a single post – I'll concentrate on only one of them here.


I Wanna Hold Your Hand

In 2006, Dr. James Coan and colleagues published a neuroimaging paper suggesting that the brains of happily married women showed an attenuation of activity related to emotion and threat when they held the hands of their husbands (Coan et al., 2006). Threat was induced experimentally by presenting a stimulus which occasionally signaled that a mild electric shock would be delivered to the ankle (20% of the time). Holding the hand of a male stranger also attenuated the hemodynamic response in some of these regions, relative to a no hand-holding control condition.²

Backing up a bit, the participants in the study were 16 heterosexual couples who rated their marital satisfaction as at least 40 on the Satisfaction subscale of the 50 point Dyadic Adjustment Scale (DAS). Total scores on the DAS were 126 for husbands (on a 151 point scale) and 127 for wives.³

The experimental design is illustrated below. The red X indicated a 20% chance of shock.

Figure adapted from a 2011 presentation by Coan (PDF), part of which can be viewed here.

At the end of each block, the women rated their subjective levels of unpleasantness and arousal on a 5 point scale. The results of the hand-holding manipulation were a bit weak. Unpleasantness ratings in the husband-hand condition were indeed significantly lower than no-hand (p=.001), but only marginally so compared to the stranger-hand condition (p=.05, with p<.05 being the usual cutoff for significance). The arousal ratings for husband-hand vs. no-hand (p=.07) and stranger-hand vs. no-hand (p=.08) were not officially significant either.

This raises a question I considered in 2006: why were the wives the only ones who were scanned?

...what about married women holding their mothers' hands? married men holding their wives' hands? unmarried women holding their partners' hands? single women holding the hands of their best friends? Perhaps the authors started with the relationship that they most expected to yield significant results...

The subjective effects of spousal-handholding were not enormous in women, which might explain why we've never seen data from husbands (i.e., perhaps there were no effects on self-report and/or neural activity). The highly-touted correlations between the wife's relationship quality rating and attenuation of threat-related brain signals weren't especially impressive either: −.59, p = .02 for the left superior frontal gyrus, −.47, p = .07 (not significant) for the right anterior insula, and −.46, p = .08 (not significant) for the hypothalamus. These numbers represent the magnitude of reduction in threat-related activity when holding the husband's hand, and were interpreted to suggest that the attenuations in pain (insula) and stress (hypothalamus) were related to the strength of attachment.⁴


Emotionally Focused Therapy

This finally brings us to the recent paper by Johnson et al. (2013). They followed the imaging protocol of Coan et al. in a set of 35 married couples who were screened for relationship dissatisfaction and scanned both before and after 23 sessions of EFT couples counseling (range of 13 to 35 sessions over 3.25 to 8.75 months). On average, the couples were white Canadians 44-45 years of age, married for 17 years. In contrast to the happy couples described above (DAS scores of 127), these couples reported moderate levels of relationship distress, with DAS scores of 80-97. For various understandable reasons, only 23 couples completed pre- and post-EFT fMRI scans. Again, only the wives were scanned.

Still, retaining 23 couples over 6 months of treatment is no mean feat. However, I will again note that there is no control condition in this experiment, so we can't know whether any changes are specifically due to the treatment of interest.

According to Johnson et al. (2013), EFT is "a manualized treatment that conceptualizes relationship distress as reflecting emotional disconnection and unmet attachment needs [18]."

Session and therapy length varied depending on the couples' presenting concerns and their progression through EFT-defined therapeutic change events [18], [28]. Specifically, when a couple was deemed according to EFT guidelines to have achieved 1) “softening” – a state of vulnerability and sharing of attachment related needs between the partners [37] – and 2) “consolidation” – where the therapist works with the couple to review treatment gains – treatment was terminated.

I am not qualified to comment on EFT and will not discuss it further, beyond saying that post-therapy DAS scores were significantly increased (pre-EFT mean=81 and post-EFT mean=96) but still, on average, in the moderately distressed range. Unpleasantness and arousal ratings in the husband-hand fMRI condition were lower after EFT.

The fMRI results after EFT were.... complicated, as shown below, and involve what appears to be post-hoc reasoning in relation to initial marital strife. Percent signal change was assessed for all voxels in the ROIs that were reported by Coan et al., which is a good and unbiased method for analyzing an independent dataset.

Fig. 2 (Johnson et al., 2013). Point estimates of percent signal change graphed as a function of EFT (pre vs. post) by handholding (alone, stranger, partner) and DAS score. Point estimates were computed separately for individuals high (+1SD) and low (−1SD) in DAS. Point estimates reflect average percent signal change (threat – safe) from all voxels activated in the original Coan et al. handholding study.


But the results are a little hard to interpret for the wives with high DAS scores, who nonetheless still experienced relationship distress. The intervention had no effect on their global threat-related brain response when holding their husbands' hands. In contrast, those with lower DAS scores showed a post-EFT increase in the threat response in the no-hand condition, a large reduction for stranger-hand, and a very large reduction for husband-hand.

Next the authors moved towards analyzing specific ROIs. I'll skip the husband vs. alone comparisons because these are less relevant. Well, except I'll quote this bizarre finding (which isn't terribly relevant, just hard to explain):

Interestingly, participants with higher DAS scores were generally less active in the substantia nigra/red nucleus when holding hands with their partners relative to when alone, independent of EFT, F(1,49.5)=6.6, p=.01.

OK then. What about the husband vs. stranger comparisons? There were a number of brain areas that showed pre- to post-therapy decreases that did not differ for husband-hand vs. stranger-hand.⁵ These regions included the right insula, which was related to relationship quality in the Coan et al. (2006) study. The two regions with positive findings (i.e., threat-related reductions in husband-hand and increases in stranger-hand) are right dorsolateral prefrontal cortex (dlPFC) and left supplementary motor area (SMA). No relationships to DAS score were reported.

Fig. 5 (Johnson et al., 2013). Percent signal change (±SE) graphed as a function of EFT (pre vs. post) by handholding (stranger vs. partner) interaction effects. Row A represents activity in the supplementary motor cortex (SMG) [sic]. Row B represents activity in the right dlPFC.


What have we learned from this study, and how does it inform the practice of EFT? If we take it at face value, the one consistent finding between the two experiments is that the threat response in right dlPFC was attenuated when holding the husband's hand, relative to holding a stranger's hand. If this neural region serves to downregulate negative emotional responses expressed elsewhere (as described below), there were no downstream regions in need of downregulation:

The dlPFC in particular supports explicit, cognitive, or “reappraisal” based self-control strategies active during unpleasant emotional states [54].  ...  The relative post-EFT inactivity of the dlPFC implies further that a secure connection with an attachment figure does not help individuals to maintain equilibrium by boosting self-regulatory capabilities per se but by reducing the perception and significance of threats, thus obviating the need for self-regulation to occur [13]. 

Having some kind of autonomic measure of threat perception (e.g., skin conductance or heart rate) would be useful in verifying this hypothesis. The authors don't interpret their other major finding, a similar effect in the left SMA (a motor control region).

The final question remains unanswered: how does this study inform the practice of EFT? The authors state:

Ultimately, our handholding paradigm has provided a unique opportunity to test some of the proposed mechanisms of social support in general, and EFT in particular, all at the level of brain function, in vivo.  

But not all of their predictions were supported. In particular, to explain the changes in neural threat processing observed in the no-hand condition, they resorted to an alternate model of therapeutic change:

We predicted that EFT would not affect neural threat responding during the alone condition.  ... [But] threat-related activity during the alone condition actually increased as a function of EFT in regions such as the dACC and portions of the PFC. Increased reactivity in these regions suggests a possible cost to increasing one's dependence upon social resources: that it becomes more difficult to tolerate being alone.
...

This is not what we observed. Although positivity ratings did not change, subjective arousal actually decreased. This suggests an alternative hypothesis: that EFT either trained or motivated clients to be more effective self-regulators even when alone.  ...  Although EFT focuses strongly on interpersonal attachments and interdependence, doing so may also increase self-regulatory motivation as clients come to value fostering effective relationships in part through self-regulatory effort.

I'm not sure that I understand this formulation, or that a dissociation between behavioral self-report and dACC activity warrants a reinterpretation of EFT's therapeutic effects. Ultimately, I don't feel like a BS-fighting superhero either, because it's not clear whether Magneto has effectively corrected the misperceptions and overinterpretations that have arisen from this fMRI research.


Footnotes

¹ Not to be confused with Emotional Freedom Techniques, or “tapping”, a rather ridiculous practice that purports to manipulate the body's energy field.

² The specific neuroimaging results were a bit less straightforward and easily interpreted than this. Regions of interest (ROIs) were defined by determining which areas were activated by the red X threat compared to the safe signal in the no-hand condition. This threat response was attenuated in the husband-hand vs. no-hand condition in the ventral anterior cingulate cortex (vACC), left caudate, superior colliculus, posterior cingulate, left supramarginal gyrus, and right postcentral gyrus. The threat response was also specifically attenuated in husband-hand vs. stranger-hand only in right dorsolateral prefrontal cortex, considered a “cognitive control” area. Finally, the stranger-hand vs. no-hand comparison revealed attenuation in the same bold blue regions above.

³ However, the correlation between husbands' and wives' DAS scores was not significant. Hmm... Would knowledge of this finding create any discord?

⁴ I won't get into how those single functions were assigned to these two complex and diverse brain regions.

⁵ Johnson et al. (2013): “In the vmPFC, left NAcc, left pallidum, right insula, right pallidum, and right planum polare, main effects of EFT revealed general decreases from pre- to post- therapy in threat activation, regardless of whose hand was held.”


References

Coan JA, Schaefer HS, & Davidson RJ (2006). Lending a hand: social regulation of the neural response to threat. Psychological science, 17 (12), 1032-9 PMID: 17201784

Johnson SM, Moser MB, Beckes L, Smith A, Dalgleish T, Halchuk R, Hasselmo K, Greenman PS, Merali Z, & Coan JA (2013). Soothing the threatened brain: leveraging contact comfort with emotionally focused therapy. PloS one, 8 (11) PMID: 24278126