Time resolution

high-resolution terminology - matching measurements at high-resolution

Time resolution

Description

Time resolution in respirometric measurements is influenced by three parameters: the response time of the POS, the data sampling interval and the number of points used for flux calculation.

Abbreviation: n.a.

Reference: MiPNet19.18C DatLab Guide, MiPNet19.18B POS-Service

MitoPedia methods: Respirometry 

Technical service

Titrations and plots of oxygen flux

A disturbance of the traces of flux are caused by titrations of chemicals into the chamber. These disturbances are increasingly pronounced as the level of dissolved oxygen declines, when oxygen is added with the solvent. Oxygen is highly soluble in ethanol, hence the effect is higher with the same volume of ethanol than water. An increase in oxygen concentration would correspond to a negative flux. The trace of flux, therefore, shows a sharp negative peak after the titration. Since 40 data points are used for the calculation of flux, the peak is smoothed and extends over a prolonged period of time, depending on the selected data sampling interval.

Oxygen flux and time resolution

Question: Upon rapid changes in oxygen concentration (as seen by switching the stirrer on and off), oxygen flux (the negative time derivative) seems to lag behind and even stretches to a much longer time span than the duration of the oxygen changes themselves would suggest. Can this problem be avoided for (kinetic) applications which require a high time-resolution?

Answer: High time-resolution is necessary in various kinetic studies. Most common applications of high-resolution respirometry, however, require only information on respiratory flux over a period of time when metabolic activity is constant. Standard settings in DatLab are optimized for such applications [MiPNet12.09]. DatLab offers a simple on-line solution for improving the time-resolution, by reducing the data recording interval (standard is 2 seconds) to the minimum of 0.2 s (setting in the Oxygraph Control window; see [MiPNet12.06]. As the data recording interval is reduced, the flux appears more noisy, but represents transitions more accurately and reduces the apparent time-delay. Most routine applications aim at obtaining average flux over short periods of time (some minutes), for which application the high smoothing effect of the standard DatLab 4 settings are suitable. Off-line, DatLab 2 offers numerous options for optimizing analyses between the opposite demands on high time-resolution versus low noise of flux. While the on-line filter in DatLab 4 is fixed (Savitzky-Golay smoothing filter applied on the 40 proceeding data points), off-line options are many.

1. A de-convolution is possible on the raw signal, applying a calibrated first-order exponential time constant (simply applying the stirrer test.
2. Subsequent to time correction (signal de-convolution), 'mild' polynomial smoothing or 'strong' arithmetic smoothing may be applied on the oxygen signal, with variable number of data points (the more data points, the stronger the smoothing effect).
3. The slope can then be calculated over a variable number of data points, not merely selecting the only on-line option for preceding data points, but including leading data points (symmetrical fit) which avoids the apparent delay in the trace of flux.
4. For the specific application of oxygen kinetics, an elaborate script (macro) applies an optimized strategy for high time-resolution combined with sufficient filtering of noise.

Question: Does the time constant of the polarographic oxygen sensor change when measurements are performed at low or high oxygen pressures?

Answer: The time constant should be independent of the absolute oxygen pressure. In the scientific literature, it has been argued that at very low oxygen the time constant increases - without any evidence, and without theoretical basis. Experimental tests in the Oxygraph-2k revealed no difference of the exponential time constant of the POS measured close to air saturation or close to zero oxygen levels.

See also: Response time of the POS

Slow responding sensor