Broadband gamma-band EEG changes during magnetophosphene perception induced by 20 Hz magnetic field stimulation

This study demonstrates that magnetophosphene perception induced by suprathreshold 20 Hz magnetic field stimulation is characterized by spatially distributed, broadband gamma-band EEG activity rather than focal low-frequency oscillations, challenging classical assumptions about visual processing markers.

The Gist

The Big Idea: "Seeing" Without Light

Imagine you are sitting in a pitch-black room. Suddenly, you start seeing little flashes of light, like tiny fireflies or static on an old TV screen, even though there is no light source in the room. This phenomenon is called a magnetophosphene.

It happens when you expose your head to a specific type of magnetic field (like a very gentle, invisible magnetic wave). Your brain "sees" something that isn't actually there. Scientists have known about this for a long time, but they've always been puzzled by one thing: What is actually happening inside the brain when this happens?

The Old Theory vs. The New Discovery

The Old Guess (The "Spotlight" Theory):
For years, scientists thought that when you see a flash of light (even a fake one), your brain's "visual center" (the back of your head, called the occipital lobe) would light up like a spotlight. They expected to see a specific, rhythmic brainwave pattern (like a steady drumbeat) in that specific area, similar to what happens when you look at a real light bulb.

The New Discovery (The "City-Wide Power Surge" Theory):
This paper says: "Actually, that's not what's happening."

Instead of a single spotlight turning on in the back of the head, the brain reacts like a whole city experiencing a power surge.

• No Rhythm: There is no steady drumbeat.
• No Single Spot: The activity isn't just in the back of the head; it spreads all over the brain, including the front (where we think and plan).
• High-Frequency Static: Instead of a slow, rhythmic wave, the brain shows a burst of fast, chaotic, high-speed energy (called "gamma-band" activity).

How They Found This Out

The researchers acted like detectives with a very sensitive microphone (EEG) placed on the heads of 13 volunteers.

1. The Setup: They put the volunteers in a dark room and zapped them with magnetic fields of three different strengths:

   • Zero strength: Nothing happened.
   • Low strength: The magnetic field was there, but too weak to make anyone "see" anything.
   • High strength: The field was strong enough that almost everyone reported seeing the flashes.

2. The Measurement: They listened to the brain's electrical chatter. They looked specifically for the "rhythmic drumbeats" (low-frequency waves) that everyone expected to find in the back of the head.

3. The Result:

   • The Drumbeat was missing: They found almost no change in the rhythmic waves at the back of the head, even when people were seeing the flashes.
   • The Power Surge was real: When people saw the flashes, the brain showed a massive, widespread increase in fast, high-frequency noise. It was like the whole brain was revving its engine, not just the visual part.

Why This Matters: The "Radio Station" Analogy

Think of your brain like a radio station.

• Normal Vision: When you look at a real object, your brain tunes into a specific, clear frequency (like a specific radio station playing a song). It's focused and rhythmic.
• Magnetic Vision: When the magnetic field triggers a phosphene, it's not like tuning into a song. It's more like static interference or a sudden burst of white noise that covers the whole dial. It's a "broadband" signal.

The paper suggests that when our brain creates a visual experience out of thin air (without real light), it doesn't use the same "focused" machinery it uses for real vision. Instead, it uses a global, state-dependent switch. It's as if the whole brain shifts into a "high-alert" mode, creating a buzz of activity that allows the perception to happen.

The Takeaway

This study challenges the old idea that "seeing" always means "lighting up the back of the brain."

• Old View: Seeing = A specific spot in the brain lights up with a rhythm.
• New View: Seeing (under magnetic stimulation) = The whole brain gets a high-speed, chaotic energy boost.

It tells us that the brain is more flexible and complex than we thought. When it creates a visual hallucination or a "ghost" image, it doesn't just turn on a single switch; it changes the entire atmosphere of the brain's electrical activity.

In short: The brain doesn't just "see" the magnetic flash; it gets excited all over, creating a high-speed buzz that lets you perceive something that isn't there.

Technical Summary

1. Problem Statement

Magnetophosphenes are visual percepts induced by extremely low-frequency magnetic fields (ELF-MF; <300 Hz). While behavioral thresholds for these percepts are well-characterized (peaking around 20 Hz), their neural correlates remain poorly understood.

• The Gap: Classical visual neuroscience suggests that visual perception is linked to focal low-frequency oscillations (specifically occipital alpha power, 8–12 Hz). However, previous EEG studies on magnetophosphenes have failed to identify a robust, reproducible focal occipital signature.
• The Hypothesis: The authors propose that magnetophosphene perception, which arises without structured optical input and likely originates in the retina, may not follow classical visual paradigms. Instead, they hypothesize that perception is associated with distributed broadband high-frequency (gamma-band) activity rather than focal low-frequency oscillatory changes.

2. Methodology

Participants and Experimental Design

• Subjects: 14 healthy volunteers were recruited; 13 were included in the final analysis after excluding one based on global data quality criteria.
• Stimulation: Transcranial Alternating Magnetic Stimulation (tAMS) was applied using a global-head coil configuration.
• Conditions: Three intensity levels were analyzed:
  1. Control: 0 mT (no field).
  2. Subthreshold: 5 mT (below perceptual threshold).
  3. Suprathreshold: 50 mT (above perceptual threshold).
• Protocol: Participants sat in darkness with eyes closed. They received 55 trials (11 intensities × 5 repetitions) of 5-second, 20 Hz sinusoidal magnetic fields. After each trial, they provided a binary behavioral report (perceived vs. not perceived).

EEG Acquisition and Preprocessing

• Recording: 64-channel EEG at 10 kHz sampling rate (to capture stimulation artifacts) using a Neuroscan system.
• Preprocessing:
  • Re-referenced to mastoids (M1/M2).
  • Epochs (5s) were extracted and time-locked to stimulation onset.
  • Filtering: Band-pass filtered (30–80 Hz) to isolate the gamma range. Notch filters removed power line artifacts (40, 50, 60 Hz).
  • Downsampling: Reduced to 2 kHz after filtering.
  • Artifact Control: Online ocular artifact reduction was used. Crucially, a subject-level quality control procedure was implemented to exclude recordings with globally contaminated signals (high RMS of difference maps) rather than rejecting individual channels.

Spectral Analysis

• Power Spectral Density (PSD): Estimated using Welch's method.
• Aperiodic-Periodic Decomposition: To distinguish true oscillatory peaks from broadband background shifts, the authors used two complementary methods:
  1. Specparam (FOOOF): Fits a 1/f background and identifies Gaussian peaks.
  2. Log-Log Linear Regression: Subtracts the linear fit from the log-log spectrum to obtain residual (excess) power.
• Statistical Analysis: Linear mixed-effects models (LMM) were used to test intensity effects (0 vs. 5 mT; 0 vs. 50 mT; 5 vs. 50 mT) across electrodes, with Bonferroni correction for multiple comparisons.

3. Key Contributions

1. Shift from Alpha to Gamma: The study challenges the assumption that visual perception (even non-visual induced percepts) is primarily indexed by occipital alpha power. It demonstrates that magnetophosphene perception is better captured by high-frequency (30–80 Hz) activity.
2. Broadband vs. Narrowband: The findings indicate that the neural correlate is broadband rather than a specific narrowband oscillation. The gamma increase persists even after removing the aperiodic (1/f) background, suggesting a genuine increase in high-frequency content rather than just a spectral shift.
3. Spatial Distribution: The study maps the spatial distribution of these effects, revealing that they are widespread (frontal and parieto-occipital) rather than focal to the primary visual cortex (occipital electrodes).
4. Methodological Rigor: The application of aperiodic-adjusted metrics (Specparam and Log-Log) in the context of magnetic stimulation provides a robust framework for separating neural signals from potential artifacts and background noise.

4. Key Results

Behavioral Results

• Perception was strictly intensity-dependent:
  • 0 mT: 0% perception.
  • 5 mT: ~1.5% perception (1 trial).
  • 50 mT: ~87.7% perception.
• Percepts were described as brief, flickering, colorless, and diffuse, often involving peripheral vision.

EEG Spectral Results

• Gamma Band Elevation: The 50 mT condition showed significant increases in gamma-band power (30–80 Hz) compared to 0 mT and 5 mT.
• Broadband Nature: No consistent narrowband peak was observed across subjects. The effect was a general elevation of high-frequency content.
• Aperiodic Correction: The gamma increase persisted after correcting for the aperiodic component using both Specparam and Log-Log methods.
  • Frontal ROI: Showed consistent, spectrally broad differences (peaks around 32–37 Hz and 68–74 Hz).
  • Parieto-Occipital ROI: Showed more restricted differences (around 34–36 Hz).
• Absence of Focal Occipital Signature: Despite strong behavioral reports, primary occipital electrodes (O1, O2, Oz) did not show statistically significant effects after multiple-comparison correction. This contradicts the classical "occipital alpha" model of visual perception.
• Spatial Pattern: The strongest effects were observed over fronto-central and parieto-occipital regions, suggesting a large-scale network reconfiguration rather than a localized sensory response.

5. Significance and Implications

• Redefining Perception Markers: The study suggests that perception induced by magnetic fields (and potentially other non-visual neuromodulation) relies on large-scale, state-dependent neural dynamics rather than localized, feedforward sensory processing.
• Retinal Origin Hypothesis: The lack of a focal occipital signature supports the dosimetric evidence that the primary interaction site is the retina, not the visual cortex. The resulting cortical activity appears diffuse and non-retinotopic.
• Methodological Shift: It advocates for moving beyond classical band-power analysis (focusing on alpha/beta) toward broadband high-frequency analysis and aperiodic-adjusted metrics when studying non-standard visual paradigms.
• Caution on Artifacts: While the authors acknowledge that high-frequency EEG is susceptible to muscle artifacts, the structured spatial distribution (frontal/parietal vs. global) and reproducibility across spectral methods argue against a trivial artifact explanation.

Conclusion: Magnetophosphene perception is not captured by classical focal low-frequency markers. Instead, it is associated with distributed, broadband high-frequency EEG changes, reflecting a global reconfiguration of cortical state rather than a specific oscillatory rhythm in the visual cortex.