Bug in PSD Plotting: Incorrect dB Scaling for Amplitude in MNE

[DingkunLiu]

Description

When plotting the Power Spectral Density (PSD) using mne.time_frequency.Spectrum.plot() or related functions with dB=True and amplitude=True, the decibel scaling appears to be incorrect. This issue arises because the conversion between power and amplitude dB follows different logarithmic relationships, and amplitude requires a factor of 20 in the scaling, whereas power uses a factor of 10 (implement here).

Example to reproduce

import mne
import numpy as np

# Generate sample data
sfreq = 1000  # Sampling frequency
times = np.linspace(0, 2, 2 * sfreq)
data = np.sin(2 * np.pi * 10 * times)  # 10 Hz sine wave
info = mne.create_info(ch_names=['sin_wave'], sfreq=sfreq, ch_types='eeg')
raw = mne.io.RawArray([data], info).load_data()

# Plot PSD with dB=True
raw.plot_psd(dB=True, amplitude=True)

Expected result

The peak at 10Hz is expected to be $120=20 * log(10^6)$, but the plot returns 60.

[welcome]

Hello! 👋 Thanks for opening your first issue here! ❤️ We will try to get back to you soon. 🚴

[drammock]

thanks for the very clear example @DingkunLiu! Fix is in #13036