Technology progressions have made it relatively easy to study physiological parameters that previously required complicated and time-consuming methods. Technology advancements include new methods to measure physiological parameters as well as advanced analysis and statistical tools (Phatak et al. 2021). Importantly, the usefulness of physiological data still depends on the quality of the recorded physiological signal and the relevance of the measured parameter.

To address these issues, we have launched a Topical Collection on Methods in Applied Physiology (https://link.springer.com/collections/gghifechbb). The EJAP welcomes original research, state-of-the-art reviews and narrative reviews in this area. Original research that compares novel and existing “gold standard” methods are appreciated. Reviews should take the approach of how new technology builds on traditional approaches and provides insight into physiological mechanisms. Animal studies that clearly demonstrate translation to human situations are welcomed as well. Please note that this Topical Collection will also include several papers already published in EJAP in the past few years. It is our hope that you find this Topical Collection useful in your work.

Some examples of technology advancements:

I.

Analysis of electromyography (EMG) records can be performed by sophisticated automated software (Felici and Del Vecchio 2020), including artificial intelligence (AI) with self-improving machine- and deep-learning algorithms.

II.

Metabolites, nucleic acids, proteins; and ions can be measured during exercise with non-invasive skin sensors, hence avoiding problems associated with obtaining blood samples and muscle biopsies.

III.

Smartwatches can record running distance and heart rate, and stored data can subsequently be used to assess physiological parameters, such as maximum oxygen uptake (VO2max), heart rate variability and so on

But there are concerns with the uncritical embracement of novel technologies, as exemplified by:

I.

Advanced software processing cannot correct insufficiently recorded EMG signals (low signal-to-noise ratio, inappropriate filtering) in a reliable manner. Difficulties may arise due to machine- and deep-learning algorithms being afflicted by poorly recorded EMG signals. Thus, it is essential to carefully scrutinise the basic recorded signals before putting these through advanced computerised processing.

II.

The relation between metabolites measured in sweat and physiological parameters is highly complex (Ali et al. 2023). Thus, from a physiological perspective, data obtained with skin sensors are not very informative before they are put into a confirmed physiological context.

III.

Fig. 1 shows records from a smartwatch equipped to measure heart rate at the wrist. At a stable running pace, the heart rate suddenly increases by about 15 beats per minute. It seems likely that this reflects an inaccuracy in rate detection rather than a real increase in heart rate, and parallel measurements with an established method would be required to determine whether the higher or lower rate is correct. Furthermore, subsequent calculations of fitness parameters (e.g. VO2max) depend on the heart rate and their validity is therefore questionable. An additional complication with the scientific use of such fitness data is that they are often determined with algorithms that are not made public by manufacturers.

Fig. 1

Running pace (darker blue shadowing) and heart rate (red line) measured with a smartwatch during running at a steady pace

Figures showing actual experimental records are important to assess the quality of measurements, and hence the methods used. Noteworthy, most scientific journals demand representative images of Western blots when proteins are studied, and we see no reason why this should not be the case also for original records of physiological measurements. Thus, figures in original research articles submitted to our journal should include experimental records, although exceptions are allowed provided authors can justify why such records are not appropriate.

We see a trend of using elaborate statistical analyses in the field of applied physiology. This often leads to Results sections where the physiology is downplayed and replaced by non-physiological descriptions of statistical main effects, interactions and correlations. In our opinion, this does not move the understanding applied physiology forward. Consequently, the EJAP Instructions for Authors states: “Thus, the text in the results section should be presented with a focus on physiology and the outcome of statistical analyses should generally be limited to those used to test specified hypotheses”.

We encourage submissions that highlight significant physiological relationships rather than statistical significance. Accordingly, we encourage graphs with individual data points, which are connected by lines when repeated measurements were performed. Conversely, we discourage the presentation of data as mean or median values with error bars, especially in the form of the frequently used bar graph format; this way of presenting data tends to move the emphasis from physiology to statistics. To exemplify this, Fig. 2 presents two sets of resting heart rate measurements obtained before and after a fictive intervention. In the first case, the results might be presented as: “The results show a statistically significant reduction in resting heart rate after the intervention (P < 0.001; Fig. 2A, left panel)”. The strong statistical significance combined with the usage of a narrow-scaled y-axis in the bar graph gives the impression of an important effect. However, when data are presented as individual points and with the y-axis starting at zero, it becomes clear that the effect was small (~ 1.5 beats/ minute) and the physiological importance was probably limited (Fig. 2A, right panel).

Fig. 2

Two fictive studies with measurements of resting heart rate before (Pre) and after (Post) an intervention. Data are presented as mean ± standard deviation (left panels) or as individual values (right panels)

The second set of fictive data presents an opposite scenario. These data might be presented as: “The intervention had no statistically significant effect on resting heart rate (P = 0.88; Fig. 2B, left panel)”. However, the presentation of individual data points shows another, physiologically more interesting picture with two groups, one where the rate was decreased and one where it was increased after the intervention (Fig. 2B, right panel), which would trigger further analyses or additional studies to reveal the underlying mechanisms. For instance, in an endurance training study, the diverging outcomes might be due to some individuals responding to the training with improved fitness and decreased resting heart rate, whereas the training was too intense for others resulting in “over-training” and increased resting heart rate. Thus, the two examples in Fig. 2 highlight the importance of showing individual data points and focus on physiological importance rather than uncritical usage of statistics.

To conclude, the results of carefully performed experiments using reliable methods followed by precise physiology-based analyses provide factual information and will drive the area of applied physiology forward. Conversely, results based on automated procedures, no inspection of raw data, and complex statistical analyses can be deceptive and lead us up the wrong alley.