A study published in 1991 by Harvard Medical School researchers explains how they did a post-mortem study of some people with dyslexia and found differences in the LGN a part of the visual pathway in our brain (3.1 Magno and Parvo in the LGN). (Link to an article in the New York Times )

Neurons that originate in the eye's retina travel to the LGN before traveling to the back of the brain to connect at the primary visual cortex (V1). From the primary visual cortex they fan out and make connections at a number of secondary locations in the back of the brain devoted to processing visual information.

Some of these neurons are grouped together to form a pathway. One of these pathways, made up of large-sized neurons, is called the Magnocellular, or M-pathway. The M-pathway processes the information captured by 'Rods' in the Retina (2.1 The Retina: Rods and Cones). Harvard professors Livingstone and Galaburda looked at the subdivision of the LGN that holds the cells in the M-pathway and found that the cells are smaller in people with dyslexia compared to non-dyslexics.

Conclusion: In our brain every part is interlinked, a smaller M-pathway leads to a less filtered input-stream of motion and orientation, from the eye's Retina to the Primary Visual Cortex (V1) in the back of our brain. Resulting in a weaker alignment system that makes it harder for dyslexics to have a Visual Grip (5.2 Visual Grid and the Origin of Ocular Dominance Patterns in V1) and causing Visual Stess. Dyslexia related deficits are also found in the Hippocampus where stress can create memory problems. (11.3 DCDC2 & Dyslexia).