When imaging it is important that all analogue signal from the imaged object is captured into a digital representation. The Nyquist rate describes the ideal sampling rate by which the image can be represented digitally without losing information. The Nyquist rate is determined as twice the highest frequency occuring in the object. It is very important for the quality of a deconvolution result that all information generated by the optics of the microscope is properly captured in digital form. Sampling rates larger than the PSF will strongly limit the deconvolution result.

We suggest using a sampling rate as close to the Nyquist rate as you can. You can calculate the Nyquist rate for your experiment by using the Nyquist calculator.
Apart from practical problems like bleaching, acquisition time and data size there is no objection at all against using a higher sampling rate than the nyquist rate (over sampling). We often suggest capturing at a lower signal to noise ratio, as the increase in quality after deconvolution with proper sampling distance but lower signal is often better.

To obtain a high quality deconvolution result, the sampling rate used to acquire the image should be as close the the Nyquist rate as possible. The relative Nyquist rate is calculated as the factor by which the actual sampling rate during acquisition exceeds the Nyquist rate (< 1 means undersampling). In order to perform deconvolution, the imaged object should be sampled at least at the scale of the point-spread-function. If the sampling rate is smaller than the PSF, deconvolution may be useless. However, slightly under sampled images can still produce acceptable deconvolution results. The minium relative Nyquist rate for which deconvolution is still reliable is 0,66 for widefield and 0,59 for confocal microscopy.