Hi Vania,

I still maintain that the first step should be to think about what kinds of computational resources your university provides or has access to before assuming that you will need to do all your image processing on your laptop. It is quite common now to run intensive processing on dedicated servers, because datasets and computational complexity are growing and it is more cost effective to consolidate and share computational power, rather than each researcher building very strong computers for themselves.

That said: yes, if you want to do image processing, and you need to do it on a laptop, you should get one with plenty of RAM (16 GB minimum), a fast processor, and a decent graphics card so that you can utilize GPU accelerated software implementations in these software packages. The Macbook Pro is probably the way to go. There isn't a lot of selection in the Mac ecosystem these days, in terms of models. You can build a Pro that is quite powerful, and that should get you through.

I hope others will back me up on this, but doing data analysis is no longer a thing one can expect to accomplish on their own workstation, let along their laptop. When projects are small, sure, you can do it on your laptop. But once you are experimenting with more complex tools, that may need to be tuned to your dataset, and require cross validation and permutation testing... or you are trying to do probabilistic network discovery... really, most things that are happening at scale are being accomplished by leveraging shared computational facilities.

Working with a shared resource has several advantages, that are actually compatible with your needs:

1. The data are hosted on a server you can access from anywhere you can install a VPN client and a SSH client.

2. Analysis can proceed 24/7 without occupying your main workstation, allowing you to travel without delaying analysis and keep your laptop with you at all times.

3. Someone else is keeping the computer working and software up to date.

4. There may be natural ways to buy in to data backups, that take place whether you feel like being responsible about backing things up or not.
So, you never have to be far from your work, even when travelling, and you can always download the bits that you need and visualize them on your laptop. I'm not saying you should buy a cheap laptop, or that performance doesn't matter at all, but I think that if you are in a mind-set where your laptop is going to be your primary analysis engine, it will not be long before you hit computational bottlenecks.

Of course, this all hinges on whether you have access to this sort of a shared computing infrastructure, how good your access to internet is, and various other things that I am taking for granted.

So, long story short, if you need to do all your work on your laptop, get the most powerful thing you can afford, buy one or two external hard drives for backup. Having lots of RAM is helpful. Maxing out the CPU is probably least important (spending hundreds more dollars for a couple extra Mhz is not good value for money). Having a fast hard drive in the system is going to make your life more comfortable if you are going to need to load large datasets into memory routinely. You will want a GPU. But before you take that approach, I would really encourage you to see what other computing options you have available, and what it would take to get remote access to them.