It depends on how HRV is evaluated. Normalized spectral components (e.g. LF/HF) may indeed reflect autonomic balance (plenty of evidence on this). However, time-domain measures (RR variance, SDNN etc), including absolute spectral power (variance again) are heavily contaminated by mean heart rate (Rocchetti et al 2000, Zaza & Lombardi 2001, Monfredi et al 2014, da Silva et al 2015).  I don't know about non linear domain, I'll look up.

Especially HRV represents the imbalances between sympathetic and parasympathetic systems.

Vegetative dysfunctions have been described for the majority of neurodegeneratives diseases in the past decades as well as depressive disorders. Nevertheless, confounders may be vegetative neuropathy due to diabetes mellitus, and some other frequent metabolic conditions as well as stroke, especially when touching the insult, anterior cingulate cortex, and quite a lot of brainstem areae. So I presume that HRV may be have to used with caution in patient populations.

Dear Dirk with regarding to stroke and other comorbid diseases to mood disorders don't you think that the cause of stroke could be an inbalanced autonomous system (multiorgan dysfunction theory) that over long time could lead to secondary diseases but that are not anymore distinguishable from the primary origin (https://www.diabetes.org.uk/Research/Our-research-projects/London/Nervous-system-imbalance-and-Type-2/)?

Dear Dick, in our opinion the minimal energy consumption principle may be acceptable for research HRV and BP interaction. We used this principle in our cognitive-hemodynamic model. If you interested in it we will be glad to invite you to our WEBinar this autumn

I would incorporate the variability in Initial Systolic Time Interval (ISTI) as a marker for the time difference between electrical and mechanical activity of the heart. This time interval is expected to be dominated by the sympathetic nerve system. The respiratory variation in ISTI has been shown to be distorted in patients suffering from Parkinson's disease. ISTI has also been shown to be sensitive to breathing stimuli.

Meijer JH, Boesveldt S, Elbertse E, Berendse HW (2008); Method to measure autonomic control of cardiac function using time interval parameters from impedance cardiography. Physiol. Meas. 29 (2008) S383-S391; http://iopscience.iop.org/0967-3334/29/6/S32/;  doi:10.1088/0967-3334/29/6/S32.

Meijer, Jan H; Hoekstra, Femke; Habers, Esther; Verdaasdonk, Ruud M; Janssen, Thomas WJ (2010). Dynamic Response of the Initial Systolic Time Interval to a Breathing Stimulus measured with Impedance Cardiography. J. Phys.: Conf. Ser. 224 012119; http://iopscience.iop.org/1742-6596/224/1/012119/; http://dx.doi.org/10.1088/1742-6596/224/1/012119.

Meijer Jan H, Elbertse Eskeline, Boesveldt Sanne, Berendse Henk W and Verdaasdonk Rudolph M (2011). Using the Initial Systolic Time Interval to assess cardiac autonomic nervous function in Parkinson’s disease. J. Electr. Bioimp. 2: 98-101; http://dx.doi.org/10.5617/jeb.216;  https://www.journals.uio.no/index.php/bioimpedance/article/viewFile/216/252.

A very good question, considering the claims that are often made for distorted HRV. It is the case that a great many conditions are characterized by modest or even extreme changes in HRV; some, like Parkinson's Disease or diabetes have been noted above, and risks for developmental conditions, like Sudden Infant Death Syndrome, could be added to the mixture. Detection of a marked loss or change in variability is very much a cause for concern. The problem comes in identifying the mechanism(s) underlying the variability change, whether the change results from some pharmacologic or other blockade at the periphery (or centrally), some process that has affected insular, cingulate or other structures that modify overall sympathetic or vagal outflow, cerebellar injury that is limiting the integration of thoracic stretch receptors with brain stem autonomic regulatory areas- anybody could go on at length for a series of disease processes. Thus, the variability issue is a very prominent sign; the difficulty is pointing to the particular mechanism that is operating.

Dear colleague,

you need to pay attention to the range VLF of the HRV spectrum. Try to start the search with Article:

Hadase M. et al. Very low frequency power of heart rate variability is a powerful predictor of clinical prognosis in patients with congestive heart failure.Circ J. 2004 Apr;68(4):343-7.

http://www.ncbi.nlm.nih.gov/pubmed/15056832

good morning,

I guess the random factor is so unpredictable, to build on it, such  formula.

I liken HRV to a sensitive but nonspecific test with many confounding variables. I am doubtful it will ever become mainstream clinical medicine for that reason. Love to be proven wrong. I'll check back on this in 40 years time. 

Would then a number of other markers if taken together indicate more as shown by Cerruti 2009?

Dear all,

As far as my experience goes, about all tests are limited in sensitivity and specificity because humans are built from an infinite number of confounding variables. That is why you need several independent tests. If you have a test with 70% specificity you will score 30% false positives. If you have a second independent test with 20% false positives (80% specificity), the combination of both tests will result in 6% false positives and therefore will have a specificity of 94%. The more independent tests, the better the diagnosis, although each test itself may have a limited sensitivity and specificity.

That is why I suggested to combine the HRV tests with the Initial Systolic Time Interval.

Dr. Tzeng, I hope to meet you over 40 years!

Dear Dr. Stefano Gaburro,

Dear All,

I used non linear domain analysis of HRV to determine the presence of cardiac dysautonomia in diabetic rodent models.

One of the most importan aspects of the HRV interpretation, is the analytic approach: spectral or fractal; I found more useful the HRV spectral analysis in terms of cardio neuropathy.

Furthermore, I consider that some variables, that are disregarded in many studies, could decrease the reliability of HRV marker :

-Ultradian and Circadian rhythms

-Infradian rhythm

-Temperature

-Age

-Gender

I would really appreciate if you could share how do you deal with this variables.

Dear Jose,

I invite you to read the publication of Prof. Oliver Stiedl in this regards considering the cyrcadian rythm but also ECG stability since he linear approach (e.g. frequency domain implies that) but most often the linear assumption are not statistically true. Therefore, non linear (fractal) or other model should be apply.

For more questions, do not hesitate to write back.

Dear all,

Please allow me to share my experience on that. Agree with previous comments and I agree the HRV reflects the brain-heart interaction as part of global ANS loop but this signal is the starting point, in my opinion.

Now, what's important, in my view, is what kind of markers we are able to extract from HRV in order to get the meaningful information? Huge amount of studies have proven correlations between physiological/pathological states and traditional LF/HF/RMSSD... markers, statistically speaking. But researchers agree as well that various limitations of the traditional methods method (5 minutes time limitation for RR pseudo-stationarity, "crafted in stone" frequency bands...) limit also the human comprehension, But the method is there and it' useful.  

One way we took to move forward is to say: ok, we have statistically averaged markers but can we also figure out what happens in real time during the experiment? In other words, we try to figure out both physiological and emotional state of the subject and to compare with the real time markers we have recorded. This also implies novel (non linear) mathematical analysis and derived markers; and the aim is not only to find new markers but also to clearly link them to the ANS activity in both cases - real time and averaged (and in a manner that anyone can check fast).     

One example is the attached experiment we have conducted - HRV/ANS monitoring during meditation, which raises a few questions (also posted in another thread).