Welcome to the Vision Journal club (VJC) 

For questions about VJC events, contact ionefine uw.edu

Student responsibilities

• All students should read the paper(s) before the paper discussion
  
• The day before the journal club all students will be expected to post one question or discussion comment about the paper - this should be submitted as an assignment on Canvas on Tuesday.
  
• Each week one student will act as discussion leader.

Class format:

Jan 3-17 we are focusing on "A hundred duck sized horses, or one horse sized duck? A discussion of sample sizes and replicability".

It's a big reading list you can find either in a zip archive, or individual files.

Jan 3 is a reading week

Jan 10, 17 are a Zoom discussion with guest discussants:  Ella Striem-Amit, Ariel Rokem, Scott Marek, Noah Benson, Steve Engel Brenden Tervo-Clemmens

Jan 24 is a student only discussion hour.

We then move to a more typical format ....

For guest lecturers we'll begin with 10 min introduction. 5 min about their career, 5 min what they think is cool or interesting about the paper. There will then be 20 min where faculty are allowed to ask questions, then faculty have to leave, and discussion continues guided by the submitted questions.

For paper discussions we'll begin with the student doing a 10 min introduction providing context about the paper - why they think it is important. This should not descibe the content of the paper because it is assumed that people have read the paper. We then go into the discussion section, based on the submitted questions.

Tues Jan 31st NO CLASS

 

Tues Feb 7th (IRL):  Maureen Neitz.

Insight from OPN1LW Gene Haplotypes into the Cause and Prevention of Myopia

Nearsightedness (myopia) is a global health problem of staggering proportions that has driven the hunt for environmental and genetic risk factors in hopes of gaining insight into the underlying mechanism and providing new avenues of intervention. Myopia is the dominant risk factor for leading causes of blindness, including myopic maculopathy and retinal detachment. The fundamental defect in myopia—an excessively elongated eyeball—causes blurry distance vision that is correctable with lenses or surgery, but the risk of blindness remains. Haplotypes of the long-wavelength and middle-wavelength cone opsin genes (OPN1LW and OPN1MW, respectively) that exhibit profound exon-3 skipping during pre-messenger RNA splicing are associated with high myopia. Cone photoreceptors expressing these haplotypes are nearly devoid of photopigment.Conversely, cones in the same retina that express non-skipping haplotypes are relatively full of photopigment. We hypothesized that abnormal contrast signals arising from adjacent cones differing in photopigment content stimulate axial elongation, and spectacles that reduce contrast may significantly slow myopia progression. We tested for an association between spherical equivalent refraction and OPN1LW haplotype in males of European ancestry as determined by long-distance PCR and Sanger sequencing and identified OPN1LW exon 3 haplotypes that increase the risk of common myopia. We also evaluated the effects of contrast-reducing spectacles lenses on myopia progression in children. The work presented here provides new insight into the cause and prevention of myopia progression.

Tues Feb 14th (IRL) Scott Murray

Weaker neural suppression in autism

Abnormal sensory processing has been observed in autism, including superior visual motiondiscrimination, but the neural basis for these sensory changes remains unknown. Leveragingwell-characterized suppressive neural circuits in the visual system, we used behavioral andfMRI tasks to demonstrate a significant reduction in neural suppression in young adults withautism spectrum disorder (ASD) compared to neurotypical controls. MR spectroscopymeasurements revealed no group differences in neurotransmitter signals. We show how acomputational model that incorporates divisive normalization, as well as narrower top-downgain (that could result, for example, from a narrower window of attention), can explain ourobservations and divergent previousfindings. Thus, weaker neural suppression is reflected invisual task performance and fMRI measures in ASD, and may be attributable to differences intop-down processing

Tues Feb 21 (IRL)  Noah Benson

Cortical magnification in human visual cortex parallels task performance around the visual field

 Human vision has striking radial asymmetries, with performance on many tasks varying sharply with stimulus polar angle. Performance is generally better on the horizontal than vertical meridian, and on the lower than upper vertical meridian, and these asymmetries decrease gradually with deviation from the vertical meridian. Here, we report cortical magnification at a fine angular resolution around the visual field. This precision enables comparisons between cortical magnification and behavior, between cortical magnification and retinal cell densities, and between cortical magnification in twin pairs. We show that cortical magnification in the human primary visual cortex, measured in 163 subjects, varies substantially around the visual field, with a pattern similar to behavior. These radial asymmetries in the cortex are larger than those found in the retina, and they are correlated between monozygotic twin pairs. These findings indicate a tight link between cortical topography and behavior, and suggest that visual field asymmetries are partly heritable.

Tues Feb 28th (IRL) Ariel Rokem

Automated Detection of Glaucoma With Interpretable Machine Learning Using Clinical Data and Multimodal Retinal Images

Purpose: To develop a multimodal model to automate glaucoma detection DESIGN: Development of a machine-learning glaucoma detection model METHODS: We selected a study cohort from the UK Biobank data set with 1193 eyes of 863 healthy subjects and 1283 eyes of 771 subjects with glaucoma. We trained a multimodal model that combines multiple deep neural nets, trained on macular optical coherence tomography volumes and color fundus photographs, with demographic and clinical data. We performed an interpretability analysis to identify features the model relied on to detect glaucoma. We determined the importance of different features in detecting glaucoma using interpretable machine learning methods. We also evaluated the model on subjects who did not have a diagnosis of glaucoma on the day of imaging but were later diagnosed (progress-to-glaucoma [PTG]).

Results: Results show that a multimodal model that combines imaging with demographic and clinical features is highly accurate (area under the curve 0.97). Interpretation of this model highlights biological features known to be related to the disease, such as age, intraocular pressure, and optic disc morphology. Our model also points to previously unknown or disputed features, such as pulmonary function and retinal outer layers. Accurate prediction in PTG highlights variables that change with progression to glaucoma-age and pulmonary function.

Conclusions: The accuracy of our model suggests distinct sources of information in each imaging modality and in the different clinical and demographic variables. Interpretable machine learning methods elucidate subject-level prediction and help uncover the factors that lead to accurate predictions, pointing to potential disease mechanisms or variables related to the disease.

 

Thurs Mar 7th (IRL) Mike Manookin

Two sides of the same coin: Efficient and predictive neural coding

Some visual properties are consistent across a wide range of environments, while other properties are more labile. The efficient coding hypothesis states that many of these regularities in the environment can be discarded from neural representations, thus allocating more of the brain’s dynamic range for properties that are likely to vary. This paradigm is less clear about how to prioritize different pieces of information that vary across visual environments. A general solution to this problem prioritizes information that can be used to predict what we will likely see in the future. Under some circumstances, the efficient coding and prediction paradigms make contradictory assertions about how the visual system in general and neural circuits in particular should behave. Here, we argue that these paradigms are complementary and often act on distinct components of the visual input. We also discuss how normative approaches to efficient coding and future prediction can be integrated.