How to prepare already preprocessed and extracted data for MNE topoplots?

My issue is the following. I have done the entire preprocessing of fNIRS data, the epoching, extraction, plotting the oxygenation/deoxygenation curves across trials, the averages across all trials, etc. My data is a dictionary containing all epoched data for all these preprocessing steps, and the inputs used to generate the plots.

Now, I want to plot the topoplots of the oxygenation during the task.
I know I could produce some average values over the oxygenation curves I made so that I have 1 value for every channel, and feed that into the plot_topomap() function together with channel locations, but if I wanted to do this for different time segments (which I do) I would need to compute the values for every section separately.

My question is: could I somehow transform the data so that plot_topomap() can compute the heatmap on its own for different time segments? I know there’s some examples on the wiki, but these use the datasets and instances of Evoked data.