Low-Frequency Local Field Potentials and Spikes in Primary Visual Cortex Convey Independent Visual Information

Andrei Belitski, Arthur Gretton, Cesare Magri, Yusuke Murayama, Marcelo A. Montemurro, Nikos K. Logothetis, Stefano PanzeriJournal of Neuroscience 2008 · DOI

Abstract

Local field potentials (LFPs) reflect subthreshold integrative processes that complement spike train measures. However, the differences in how LFPs and spikes encode rich naturalistic sensory stimuli remain poorly understood. We recorded LFPs and spikes from the primary visual cortex (V1) of anesthetized macaques during color movie presentations. We found that the most informative LFP frequency ranges were 1–8 Hz and 60–100 Hz, while spike power was informative only at frequencies <12 Hz. High-gamma LFPs (60–100 Hz) and spikes showed high signal correlations, suggesting they are generated within the same network. Conversely, LFPs <40 Hz showed minimal correlation with spikes or high-gamma LFPs, indicating that low-frequency LFPs reflect neural processes that are functionally decoupled from spiking output under natural conditions.

Introduction

A central debate in systems neuroscience is whether local field potentials (LFPs)—the aggregate rhythmic electrical activity in a brain region—are merely a "blurred" version of spiking activity or if they provide unique information about internal processing. While spikes represent a neuron's output, LFPs are thought to reflect subthreshold integrative inputs. Most previous research has used simple stimuli like bars or gratings. This study investigates how these signals behave during the processing of complex, naturalistic stimuli, specifically color movies, to determine if LFPs and spikes provide redundant or complementary information about the visual world.

Methods and Data

Neural activity was recorded from the primary visual cortex (V1) of four anesthetized rhesus macaques across 44 sites. The animals viewed 3.5 to 6-minute color movie clips, with 12–40 repetitions per movie to allow for trial-averaging and noise correlation analysis.

The researchers employed several key technical approaches:

  1. Signal Separation: Raw signals were decomposed into LFPs (1–250 Hz) and multiple-unit activity (MUA/spikes).
  2. Spectral Estimation: The multitaper technique was used to calculate the power spectrum within 2-second windows.
  3. Information Theory: Mutual Information (MI) metrics quantified the "bits" of information specific frequency bands carried about the movie.
  4. Correlation Analysis: The team distinguished between "signal correlations" (average response similarity) and "noise correlations" (shared trial-to-trial fluctuations).
Representative single-trial LFPs, spike rasters, and their corresponding instantaneous power/rates over time.

Key Insights: The Frequency-Dependent Information Profile

The study revealed that visual information in the LFP is not distributed uniformly across the spectrum. Instead, it exhibits a "two-peak" profile, with maximum information content in the low-frequency range (1–8 Hz) and the high-gamma range (60–100 Hz). Frequencies between 12–40 Hz (alpha and beta bands) carried significantly less information, potentially reflecting internal neuromodulatory states rather than external sensory features.

Average power spectra and spectrograms for LFPs and spikes during movie viewing versus spontaneous activity.

A critical finding was the functional decoupling at approximately 40 Hz. Below this threshold, LFPs showed almost no signal or noise correlation with spikes. Above 40 Hz, LFPs and spikes became highly synchronized. This suggests that high-gamma LFPs and spikes are redundant; knowing one provides little additional information if the other is already known. In contrast, low-frequency LFPs and spikes are independent. Combining these two signals nearly doubles the total extractable information about the movie.

Results and Quantitative Analysis

  • Information Capacity: Both the 1–8 Hz and 60–100 Hz LFP bands carried approximately 0.22 bits of information. Spikes were more informative at very low frequencies (~0.35 bits).
  • Additivity: The joint information of a low-frequency LFP (4 Hz) and a spike signal (1 Hz) was 0.57 bits, nearly the exact sum of their individual parts (0.22 + 0.35), confirming their independence.
  • Correlations: High-gamma LFPs (72 Hz) and spikes (76 Hz) exhibited a high signal correlation of 0.83. Conversely, the correlation between 16 Hz and 72 Hz was negligible (0.05). Low-frequency LFPs (<24 Hz) shared strong noise correlations (~0.37), suggesting a common, diffuse source of variability, such as ascending neuromodulatory pathways.

Discussion and Future Directions

These findings demonstrate that low-frequency LFPs represent a distinct channel of sensory processing in V1, likely reflecting subthreshold integrative inputs that are decoupled from local spiking output.

However, the use of anesthetized subjects is a notable limitation, as anesthesia suppresses top-down attention and higher-order feedback, which are known to influence low-frequency oscillations. Future research in awake, behaving animals is necessary to determine if this independence persists. Additionally, further "reverse-correlation" studies could help identify the specific visual features (e.g., motion, luminance) encoded by these informative low-frequency LFP fluctuations.

Summary updated

Abstract

Local field potentials (LFPs) reflect subthreshold integrative processes that complement spike train measures. However, little is yet known about the differences between how LFPs and spikes encode rich naturalistic sensory stimuli. We addressed this question by recording LFPs and spikes from the primary visual cortex of anesthetized macaques while presenting a color movie. We then determined how the power of LFPs and spikes at different frequencies represents the visual features in the movie. We found that the most informative LFP frequency ranges were 1-8 and 60-100 Hz. LFPs in the range of 12-40 Hz carried little information about the stimulus, and may primarily reflect neuromodulatory inputs. Spike power was informative only at frequencies <12 Hz. We further quantified "signal correlations" (correlations in the trial-averaged power response to different stimuli) and "noise correlations" (trial-by-trial correlations in the fluctuations around the average) of LFPs and spikes recorded from the same electrode. We found positive signal correlation between high-gamma LFPs (60-100 Hz) and spikes, as well as strong positive signal correlation within high-gamma LFPs, suggesting that high-gamma LFPs and spikes are generated within the same network. LFPs <24 Hz shared strong positive noise correlations, indicating that they are influenced by a common source, such as a diffuse neuromodulatory input. LFPs <40 Hz showed very little signal and noise correlations with LFPs >40 Hz and with spikes, suggesting that low-frequency LFPs reflect neural processes that in natural conditions are fully decoupled from those giving rise to spikes and to high-gamma LFPs.
BibTeX
@article{belitski2008low,
  author = {Andrei Belitski and Arthur Gretton and Cesare Magri and Yusuke Murayama and Marcelo A. Montemurro and Nikos K. Logothetis and Stefano Panzeri},
  title = {{Low-Frequency Local Field Potentials and Spikes in Primary Visual Cortex Convey Independent Visual Information}},
  journal = {Journal of Neuroscience},
  year = {2008},
  doi = {10.1523/jneurosci.0009-08.2008},
}
Figure