It's great to see research being done on this topic. I discussed the research more with the publication's authors.
After the publication, they performed additional testing with a wireless packet sniffer while repeatedly turning on a phone. They were not able to capture any WiFi packets even when hopping between different WiFi channels.
The radio transmissions they see during boot are only in the 2.4GHz band. They don't see any activity in the 5GHz band. A WiFi scan typically scans across both 2.4GHz and 5GHz bands. They also suspect the transmissions they observe could be too narrow band to be WiFi.
These are just observations and not necessary definitive conclusions. It's possible the radio power measurements are not caused by WiFi. Instead it's possible they're caused by Bluetooth, which only operates in the 2.4GHz band.
The AOSP code for the Bluetooth stack has lots of logging at the DEBUG log level. If someone has access to a GrapheneOS build with debug logging enabled, those debug log statements could help to determine if Bluetooth is causing radio activity during boot.
Alternatively, another test would be to sniff Bluetooth packets in an isolated environment during phone boot for someone who has access to to a device that can capture Bluetooth packets.