MRI is known to have noise in Rician distribution. However, apart from the raw data, HCP also provides data after some preprocessing step (a.k.a. minimal processing pipeline) in which some ICA-based denoising is also included. I wonder after these processing, will the Rician assumption of noise still hold?