Is it easy to change field of research when moving from PhD to postdoc?

The general answer would be no, that it isn't easy, but it isn't a yes/no situation. It depends on many things. A switch between related subfields of a general field would be easier (won't say easy) than a more radical change.

But it depends even more on what you have overall to offer to a position. If you are familiar with the research process and the nature of evidence in a field and have some experience with that, you will probably be considered, though others more closely aligned will probably have a better chance.

That said, any position you don't apply for you won't get.

I'm an outsider to your field, but your description sounds like you have some alignment. Make your case to whoever is hiring.

To get hired for a postdoc position, the PI needs to believe that you are the best candidate available to them. Most of the time they will have candidates available with directly related experience, so you need to somehow make the case that you are a stronger candidate.

In your example, maybe you can make the case that while you don't have the specific experience in graph theory and neural networks etc, your statistical background will allow you to pick up the machine learning side of the research relatively quickly, and you also bring a strong understanding of the subject matter which is crucial for building clinically relevant models, etc.

I have recently been talking with my supervisors about a similar switch in fields when I graduate. If your supervisors believe you are capable of making the switch, it will help a lot if you're able to lean on their networks. Do they have colleagues working in your target field that they can recommend you to? A prospective PI is much more likely to take a chance on a field-switching postdoc if that PI's trusted collaborator can explain why they believe the postdoc will be successful.

If you have time left before the end of your PhD you could try to publish something more related to your target field, or which showcases some relevant skills which are not otherwise demonstrated in your PhD project. This makes you a much less risky prospect. For example, maybe you could use a neural network in your analysis of MRI data. (There is some interesting work on MRI dataset harmonisation using neural networks, for example).

There is a planned move from one field to another, and then there is taking on a project out of curiosity in an area that is new to you.

A planned change of field

This is something that takes a lot of work and study, which will have to be self-directed (no-one will tell you exactly what you need to do), and which will probably be in addition to your normal duties. It is certainly possible.

You might want to spread the work out over a longer period. A typical way to go about this might be to seek a postdoc in your current field, but one with a decent amount of time for your own research. It is likely that you will need to find collaborators who already work in your target field.

A spontaneous project in a new area

This is common and much easier. It usually means a side-project, rather than a new job.

If you have a very strong research record, however, or a truly outstanding academic record in other ways, then you might well be able to secure a job in a new field on the basis that you will pick up the skills you need as you go along.

The specific example

I am in pure maths, so take this with a perfectly spherical grain of salt: I think that a postdoc in mathematical modelling probably has a rigid minimum requirement that candidates have experience, at research level, in some kind of mathematical modelling in a broad sense. You might have done something that fits the bill.