Daily Page: 2025-12-03

• L1 – Photoreceptor layer: Contains rods (low-light, peripheral vision) and cones (color, high-acuity central vision). They convert light into electrical signals via phototransduction.
• L2 – Outer nuclear layer (ONL): Cell bodies of rods and cones; houses their nuclei.
• L3 – Outer plexiform layer (OPL): Synaptic connections between photoreceptors and horizontal/bipolar cells. Photoreceptor terminals form ribbon synapses with bipolar and horizontal cells.
• L4 – Inner nuclear layer (INL): Contains nuclei of bipolar, horizontal, and amacrine cells, plus Müller glial cell bodies.
• L5 – Inner plexiform layer (IPL): Synaptic zone where bipolar cells connect to ganglion cells and amacrine cells modulate these signals. Subdivided into ON and OFF sublaminae.
• L6 – Ganglion cell layer (GCL): Cell bodies of retinal ganglion cells whose axons form the optic nerve, carrying processed visual information to the brain.

These layers enable parallel ON/OFF pathways, center-surround receptive fields, and initial feature extraction (motion, direction, color) before signals reach the visual cortex. (134 words)

• Layer 1: Mostly dendrites from deeper layers and some inhibitory neurons; carries “context” and predictive signals from higher cortical areas and the thalamus.
• Layer 2/3: Pyramidal cells that integrate local information and send predictions/outputs to other cortical columns and higher areas. Key for forming object representations in the Thousand Brains model.
• Layer 4: The main input layer in sensory cortices; receives feed-forward sensory data (from thalamus in primary areas). Contains spiny stellate cells in granular cortex.
• Layer 5: Output layer that sends signals to motor areas, brainstem, and superior colliculus. In the Thousand Brains model, L5 cells also project to the basal ganglia for voting on actions.
• Layer 6: Sends feedback to the thalamus and modulates input in Layer 4. Crucial for attention and controlling what sensory information is allowed into the column.

Hawkins uses this six-layer neocortical architecture as the universal building block for his hierarchical temporal memory (HTM) and Thousand Brains Theory, not the retina.

… is a text-based language designed to standardize the representation and communication of Active Inference generative models. It aims to enhance clarity, reproducibility, and interoperability in the field of Active Inference and cognitive modeling.