Diagnosing the Neural Circuitry of Reading Brian A. Wandell, Rosemary K. Le  Neuron  Volume 96, Issue 2, Pages 298-311 (October 2017) DOI: 10.1016/j.neuron.2017.08.007 Copyright © 2017 Elsevier Inc. Terms and Conditions

Figure 1 The Reading Circuitry Retinal photoreceptors encode the image and then transform this encoding with multiple specialized neural circuits. Retinal ganglion cells project directly to the lateral geniculate nucleus (LGN) and via the superior colliculus to the pulvinar. The parvocellular (P), magnocellular (M), and koniocellular (K) layers of the LGN project to the V1 and extrastriate cortex. The visual regions within the pulvinar project mainly to the extrastriate cortex. These thalamic nuclei also receive many inputs from the cortex. The colors overlaid on the posterior cortex show the locations of some visual field maps in the posterior cortex. Learning to read produces specialized circuitry that selects certain signals for further processing. Posterior visual signals are transmitted to the auditory and language system (red overlay, white text). Five general cortical regions contain subdivisions that are consistently identified as active during reading: the VOT, which includes the visual word form area (VWFA); regions within the intraparietal sulcus (IPS), which appear to be a source of top-down modulation; regions near the primary auditory cortex in the superior temporal gyrus (STG), where adult dyslexics have low activation while integrating letters with speech sounds; and Wernicke’s area and Broca’s area, which are implicated in the comprehension and production of language. Four large tracts (black text) terminate near the functionally defined regions and/or contain subdivisions with tissue properties that are consistently found to differ between groups of good and poor readers: the VOF, ILF, superior longitudinal fasciculus (SLF), and the AF (Arcuate). Neuron 2017 96, 298-311DOI: (10.1016/j.neuron.2017.08.007) Copyright © 2017 Elsevier Inc. Terms and Conditions

Figure 2 Pathways of Vision and Reading (A) Major tracts with one termination in the occipital lobe (left) or two terminations in the occipital lobe (right) These tracts transmit information to many places in the human brain. (B) Tissue properties of four white matter tracts are correlated with reading behavior: ILF (yellow), AF (pink), SLF (blue), and VOF (green). These pathways connect cortical regions that are responsive during reading. For clarity, we show only estimated fiber bundles (streamlines) near the core of each tract. Neuron 2017 96, 298-311DOI: (10.1016/j.neuron.2017.08.007) Copyright © 2017 Elsevier Inc. Terms and Conditions

Figure 3 Measures of Phonological Processing and Reading Are Strongly Correlated The colored spheres represent the value of three behavioral measures. The vertical axis is the score on a comprehensive test of phonological processing (CTOPP). The other two axes are the test of word reading efficiency (TOWRE) and the Word ID component of the Woodcock Johnson reading mastery test (Word ID, WJRMT). Measurements were obtained as part of a longitudinal study that sampled performance at four time points spread over 3 years. The color of the sphere represents the age of the subject at the time of test. Scatterplots between pairs of measures are projected onto the three bounding planes (gray dots). The correlation between the two reading measures is 0.79 (TOWRE and WJRMT). The correlations between the phonological measure and the reading measures are 0.62 (CTOPP and WJRMT) and 0.52 (CTOPP and TOWRE). The correlations are present at all ages. Neuron 2017 96, 298-311DOI: (10.1016/j.neuron.2017.08.007) Copyright © 2017 Elsevier Inc. Terms and Conditions

Figure 4 Quantitative Measurements from Retinotopic Maps (A) Retinotopic maps can be identified automatically within a few millimeters using either functional or anatomical scans. The estimated position of twenty maps can be derived from T1-weighted anatomical data using open-source algorithms (Benson et al., 2012, 2016; Fischl et al., 1999; Wang et al., 2014). (B) The size (degree [deg]) of the population-receptive field at 3 deg of eccentricity can be estimated using a simple fMRI measurement. Data from several laboratories, using different stimuli and methods, roughly agree. Refining and harmonizing the methods should further reduce the between-site variation. This input-referred quantitative measurement may be useful for assessing the integrity of these visual field maps in a reading-impaired subject. Neuron 2017 96, 298-311DOI: (10.1016/j.neuron.2017.08.007) Copyright © 2017 Elsevier Inc. Terms and Conditions

Figure 5 FOV of the VOT Reading Circuitry (A) The FOV of the bilateral VOT reading circuitry is shown for a representative subject. The gray dots are the pRF centers; the preponderance on the right reflects the larger number of voxels in the left hemisphere. The color indicates the relative sensitivity at each point in the visual field. The dashed line is the half-maximal (0.5) sensitivity contour. (B) The group average FOV of the bilateral VOT reading circuitry (n = 20). In most subjects, the VOT reading circuitry responds to stimuli located along the horizontal midline, extending to about 8 deg eccentricity. The dashed line is the half-maximal (0.5) sensitivity contour. This quantitative measure has the potential to serve as a diagnostic measure. Neuron 2017 96, 298-311DOI: (10.1016/j.neuron.2017.08.007) Copyright © 2017 Elsevier Inc. Terms and Conditions

Figure 6 Circuit Model of the Visual Reading Circuitry (A) The retinal image of text is processed by the visual cortex and transmitted to the VWFA region in the VOT. Under some task conditions, the cortex in the IPS modulates the VWFA signal. The VOF contains axons that carry signals between the dorsal and ventral posterior cortex; these axons are well positioned to communicate between the VWFA and the IPS. (B) In most subjects, the dorsal region within the reading circuitry is within the IPS maps (Cohen et al., 2008; Kay and Yeatman, 2017). Neuron 2017 96, 298-311DOI: (10.1016/j.neuron.2017.08.007) Copyright © 2017 Elsevier Inc. Terms and Conditions

Figure 7 Quantitative MRI Measures Coupled with Tractography Can Diagnose the Status of White Matter Tracts The x axis indicates the normalized position along a tract (100 equally spaced sample points). The y axis measures the T1 (second) relaxation time at 3 Tesla (3T) for the left SLF and the left arcuate, two tracts implicated in reading. The dark blue line is the population mean (n = 96), and the blue shaded region indicates 2 SD of the mean. T1 values from a single subject (purple) are compared with the population norm. Neuron 2017 96, 298-311DOI: (10.1016/j.neuron.2017.08.007) Copyright © 2017 Elsevier Inc. Terms and Conditions