Tutorial Contents

Align trace baselines

Constructed data

Shift traces

Transform baselines

Real data

Contents

Align Trace Baselines

Sometimes, a multi-trace recording consists of a relatively small and brief signal in each trace, superimposed on a baseline that may vary considerably between traces. Typically, this can occur in calcium-imaging data, where baseline fluorescence may vary between different regions of interest recorded simultaneously. It can be difficult to visualize such data, and thus difficult to decide what, if any, further analysis is useful, and on which regions. It can be helpful in these circumstances to eliminate the baseline shift between traces, while still viewing the various traces at the same gain, so that differences in the signal become apparent.

DataView provides two possible methods to align the baselines of multiple traces. One is to shift the traces individually up or down until each lies within its axis limits, without changing their scale. This is a purely cosmetic change that does not affect the data. The other is to apply an offset to the actual data of each trace, so that they all have the same baseline value. This changes the data themselves, and should ONLY be done if the baseline value is irrelevant, perhaps because you are only interested in changes from the baseline (and it should be properly documented in any report).

Constructed data

Clearly some traces have a sine wave, but the differences in amplitude are obscured because they all occupy the full height of their cell due to the autoscaling applied when the file was constructed.

This sets all the axes to have the same scale as axis 1. However, the different traces have different fluorescent baselines, so many are displaced outside their axis limits. In fact, the only trace now visible in the Matrix view is trace 1.

Shift traces

We could select each axis in turn and manually apply the Traces: Move up (down trace) and Move down (down trace) commands to shift each to lie within its axis range, but this would be very tedious. DataView can automate this process.

The axis scales of traces 2-16 have been adjusted so that the start of each trace now lies within its axis limits at the same relative position as trace 1, and all traces are at the same scale. However, at the gain of trace 1, many traces extend vertically beyond their display cell in the Matrix view.

The Matrix view now shows each trace positioned within its own axis. The traces are at the same gain, so amplitude comparisons can be made visually, but absolute DC level of each trace is different: they have been vertically-shifted to bring them into view. Trace 1 has a low amplitude signal, trace 2 has a large signal, trace 3 has a medium signal, and trace 7 has no signal at all.

It might be nice to view the traces superimposed on each other to see the differences more clearly.

The Cycle trace view displays all the traces on the same axis, with the scale settings taken from axis 1. However, the real differences in baseline between the traces means that many are out of range in the Cycle view. It is time to try the other adjustment.

Transform baselines

In the new file, the data in traces 2-16 have been offset so that their average value between time 0 and 50 is the same as that in trace 1 (the Master trace).

When the new file loads the Cycle Traces and Matrix views automatically switch to use it as the source for data. The Cycle traces view looks good, but some traces are off scale in the Matrix view. This is because the main view has retained its previous axis scale settings.

REMEMBER: you have permanently changed the data in traces 2-16 in the new file, so they are no longer what was recorded in the original experiment (or, in this case, constructed from equations).

Real Data

Examples of using this facility with real data can be found in tutorials on removing photo-bleaching effects and evoked potentials.