Informative features of local field potential signals in primary visual cortex during natural image stimulation

Area of Science:

• Neurophysiology
• Computational Neuroscience
• Visual Neuroscience

Background:

• Local field potentials (LFPs) are crucial for understanding network activity and brain-machine interfaces.
• Uncertainties exist regarding the type and extent of visual information encoded in LFPs, especially under naturalistic viewing conditions.

Purpose of the Study:

• To investigate the visual information content of LFPs in response to natural images.
• To determine the specific LFP features and temporal dynamics that convey visual information.
• To assess the spatial scales at which LFPs encode visual stimuli in the primary visual cortex (V1).

Main Methods:

• Recorded LFP responses to natural images in awake and anesthetized macaques using Utah multielectrode arrays.
• Trained Gabor wavelet (GW) models to identify presented images from LFP data.
• Analyzed LFP features, temporal dynamics, spectral bands, and spatial scales for image discriminability.

Main Results:

• Natural images were successfully identified from LFP responses using GW models, indicating shared information with spiking activity.
• Simple scalar LFP metrics provided robust image identification, highlighting the informativeness of basic LFP features.
• Image identification was most accurate within the first 300 ms post-presentation, with information present across all spectral bands.
• Optimal image identification occurred at spatial scales of ~0.5° and with multiple orientations, suggesting sensitivity to fine-grained visual details.

Conclusions:

• LFPs contain significant visual information about natural images, comparable to spiking activity.
• Short-latency, broadband LFP responses are critical for encoding visual information during natural scene viewing.
• LFPs encode stimulus-specific visual information at fine spatial scales within macaque V1, relevant for both basic research and brain-machine interface applications.