General SSP crash - memory related?

Hello,

I've been preprocessing some MEG recordings, and I'm trying to run a general SSP on the empty-room noise recording to try and correct for a couple of wonky sensors, but this operation causes my computer to completely freeze, such that I was forced to hard-boot it. Because of this, I unfortunately was not able to obtain an error message, but I believe the problem is just that my computer doesn't have the resources to handle this operation. I'm hoping that someone can take a look at the details below and tell me whether that makes sense, or whether there could be something else going wrong.

RECORDING:
Channels: 297; duration: 120s; Sampling: 2400Hz;
conditions: empty room (noise)
Preprocessing: notch filters at 60, 120, 180Hz, Band pass: 0.3-40Hz

SOFTWARE:
Brainstorm 25-Jun-2020
MATLAB R2020a

COMPUTER SPECIFICATIONS
model: Lenovo ideapad 320-15IKB Touch
Processor: Intel Core i3-7100U @2.40GHz
Installed RAM: 6.00GB (5.88 usable)
System type: 64-bit OS, x64-based processor
Free space on hard-drive: 659GB
OS: Windows 10 Home, Version 1909

PROCESS:
SSP: Generic;
time: ALL; event name: (empty/continuous); Frequency band: 5-50Hz;
Sensor types: MEG;
Compute using existing SSP: YES
Save averaged artifact: NO
Method to calculate projectors: PCA 1 component/second

NB: these exact same settings functioned appropriately when I reduced the time window to 60s, and took ~10 minutes, but caused noticeable slowdown for my computer until it finished.

ERROR DESCRIPTION:
After hitting "run", the progress bar for "processing recordings" appears, stays at 0% for ~1min, then abruptly reaches 100%, at which point my computer freezes completely. I left my computer for ~1hour to see if there was any change, but there was not, and I was forced to hard-boot. The crash occurs reliably every time I have tried this operation. Manually tagging the wonky-looking sensors as Bad before running the SSP does not prevent the crash.

Let me know if there's any other information that could help pinpoint the error, and thanks for your help,

-CJ

Well, it looks like you already find your: your system crashes when Matlab uses too much memory.
A windows home laptop with 6Gb RAM seems indeed tight to work with MEG signal processing.

We didn't design the SSP computation to work on large data segments. It is not optimized for this use.
If you can't work on a computer with more memory, just use a shorter time window, it should not change the results at all.

The fact that it freezes your computer is not a Brainstorm bug, but a problem of safety of your operating system. Your OS should not accept allocating amounts of resources (memory, swap or processor) to a level that would crash the system.

If you know that a few sensors are bad, why don't you simply mark them as bad instead of going through this convoluted solution?
https://neuroimage.usc.edu/brainstorm/Tutorials/BadChannels