Hi,

Try to observe the high and low frequency components (HF and LF) of HRV.

Also, this discussion could be interesting for you:

https://www.researchgate.net/post/Can_anyone_explain_why_HRV_is_reduced_during_attention

Best regards

A puzzle it is, the way you describe it. The only thing that occurs to me is that a degree of "mindfulness" -- objective self-awareness that allows them to disclose negative emotion -- mitigates the physiological effect of the negative emotion. Self-report isn't the same as the actual experience of an emotion. Those who minimize their distress in reporting may be less accepting of it, more ashamed, overcome momentarily with stressful memories of speaking in public, etc.

Just a guess, but if you interviewed some of the subjects about what went into their rating and report of negative emotion, you may find it's more complex than expected.

Also, you could try monitoring GSR along with HRV and see how well they correlate , and then check which correlates better with the self-report. These basic studies are important or have not been reported, but they get to the heart of subjective experience and its influence on the body variables.

Good luck!

Hi Salome,

In my experience HRV measurements during speech are very difficult to interpret: speaking is really a series of Valsalva-like maneuvers. If you want to extract data related to anxiety or stress, look at silent periods, rather than during speech. Better still, you might follow blood pressure levels by Finapres, and look for signs of stress there.

As John indicated, interpretation of HRV during speech is difficult. Respiration significantly influences HRV and speech involves alteration in respiration.

One other very important point is to not look for all the HRV indexes to see which one gives you significant P value! this often happens in psychophyiological studies using HRV. Make a clear hypothesis, select the right HRV index, and then stick to that. If you would like to look at all indexes that's fine but report them all and interpret considering all.

Also, try to analyze your data as within-subject better than between-subject. If all the subjects experienced different emotions then within-subject comparison is more appropriate with HRV since there is huge inter-individual variability.

I very much agree with both John and Ali: RSA is a phenomenon which is dependent upon respiration rate and pattern. Speech drastically alters both and it is therefore very difficult to estimate RSA and make sense of changes related to RSA under other conditions. I would say one learns nothing or very little about cardiac prasympathetic control with such results. I agree with John that you should compare quiet periods--not speaking ones (also speaking often slows down the breathing rate so that your real repiratory band may be below the frequency range you are using--often .15-.50 Hz). Also as John and I demomstrated together some 27 years ago: not only does RSA change independently of cardiac vagal tone when respiration rate is altered but that respiration rate does vary alot in different (behavioral) situations. So your correlations are as (or actually more) likely to be a function of respiration rate as to be a function of anything vagal happening to the heart.

I would suggest you simply add heart rate changes to the variables you are examining (and blood pressure too)--each directly reflecting changes in autonomic tone. I also very much agree with Ali that you should stick to one measure of HRV to prevent spurious results.

Dear colleagues,

Thank you all so much for your thoughtful and helpful comments! All of them are very clear to me.

I actually had split up my Task into three episodes: Listening to Instructions, Thinking and preparation (silent), and Speaking. So I have HRV (and other) indices for three task conditions, I could say.

I went back to check my data and what I find is that despite the overall direction of change in absolute levels (i.e. increase or decrease) my Delta scores (for both RSA and RMSSD) remain negatively correlated with self-reported negative emotions across conditions. For RSA, in Listening to Instructions condition, correlations are not significant, but negative; in Thinking and preparation condition – they are negative and marginally significant (p < 0.1); in Speaking condition they are negative and significant (p < 0.05). RMSSD exhibits very similar pattern of relationships with emotions.

So, I think I can conclude that negative emotion experience is associated with lower Deltas despite the change direction in absolute levels?

I have respiration rate, hear rate and skin conductance data for the same periods. Respiration rate increases while listening to Instructions, but later slightly decreases in Preparation stage, and falls significantly below baseline in speech (as suggested above). Heart rate increases in all three conditions. As for skin conductance across three conditions, I observe increasing arousal as measured with SCL and frequency of NS SCRs. I don’t find any relationship between reported emotions and skin conductance change though. Sympatho-vagal ratio (low to high frequency ratio) also steadily and significantly increases from Instructions to Preparation and to Speech. I think I can conclude based on this pattern of changes that sympathetic tone increases.

P.S. this was my first ever question on research gate, and I am not very familiar with the "local" etiquette rules, so excuse my long response :)

thanks again!