Description of the problem

the mne function psd_array_welch uses the scipy.signal.spectrogram function for each segment. The latter function has the option detrend which is set as default to 'constant'. This signifies that the mean is subtracted when calculating the psd, which is most probably not desired and not documented in mne. I do not believe it would make sense as a default behavior. I suggest that the detrend option be set to "False" when calling spectrogram from mne, or at least given a choice.

Steps to reproduce

import mne
import numpy as np
import matplotlib.pyplot as plt
import scipy.fft as sfft

n = 256
x = np.random.rand(n)-0.4 # not centered
fs = 100
psd1, f = mne.time_frequency.psd_array_welch(x, fs, window='boxcar')
dft = sfft.rfft(x)
psd2 = np.abs(dft * np.conjugate(dft)) / n /fs
psd2[1:] *= 2

plt.figure()
plt.plot(f, psd1, label='mne')
plt.plot(f, psd2, label='mine')
plt.legend()
plt.show()

print(psd1[:5])
print(psd2[:5])

output:
[1.50992908e-33 2.62355362e-03 1.04821458e-04 1.20104558e-03
 7.80199506e-04]
[0.03302601 0.00262355 0.00010482 0.00120105 0.0007802 ]

Link to data

No response

Expected results

psd1[0] should be the same as psd2[0]

Actual results

psd1[0] and psd2[0] are different

Additional information

Platform macOS-13.3.1-arm64-arm-64bit
Python 3.10.11 | packaged by conda-forge | (main, May 10 2023, 19:01:19) [Clang 14.0.6 ]
Executable /Users/chapon01/Programs/miniconda3/envs/mne/bin/python
CPU arm (12 cores)
Memory 32.0 GB

Core
├☑ mne 1.4.0
├☑ numpy 1.23.5 (OpenBLAS 0.3.21 with 12 threads)
├☑ scipy 1.10.1
├☑ matplotlib 3.7.1 (backend=module://matplotlib_inline.backend_inline)
├☑ pooch 1.7.0
└☑ jinja2 3.1.2

Numerical (optional)
├☑ sklearn 1.2.2
├☑ numba 0.56.4
├☑ nibabel 5.1.0
├☑ nilearn 0.10.1
├☑ dipy 1.7.0
├☑ openmeeg 2.5.6
├☑ pandas 2.0.1
└☐ unavailable cupy

Visualization (optional)
├☑ pyvista 0.39.0 (OpenGL 4.1 Metal - 83.1 via Apple M2 Pro)
├☑ pyvistaqt 0.0.0
├☑ ipyvtklink 0.2.2
├☑ vtk 9.2.6
├☑ qtpy 2.3.1 (PyQt5=5.15.6)
├☑ pyqtgraph 0.13.3
├☑ mne-qt-browser 0.0.0
└☐ unavailable ipympl

Ecosystem (optional)
└☐ unavailable mne-bids, mne-nirs, mne-features, mne-connectivity, mne-icalabel, mne-bids-pipeline