Ariel Rokem

Dipy, a library for the analysis of diffusion MRI data

Diffusion Imaging in Python (Dipy) is a free and open source software project for the analysis of data from diffusion magnetic resonance imaging (dMRI) experiments. dMRI is an application of MRI that can be used to measure structural features of brain white matter. Many methods have been developed to use dMRI data to model the local configuration of white matter nerve fiber bundles and infer the trajectory of bundles connecting different parts of the brain. Dipy gathers implementations of many different methods in dMRI, including: diffusion signal pre-processing; reconstruction of diffusion distributions in individual voxels; fiber tractography and fiber track post-processing, analysis and visualization. Dipy aims to provide transparent implementations for all the different steps of dMRI analysis with a uniform programming interface. We have implemented classical signal reconstruction techniques, such as the diffusion tensor model and deterministic fiber tractography. In addition, cutting edge novel reconstruction techniques are implemented, such as constrained spherical deconvolution and diffusion spectrum imaging (DSI) with deconvolution, as well as methods for probabilistic tracking and original methods for tractography clustering. Many additional utility functions are provided to calculate various statistics, informative visualizations, as well as file-handling routines to assist in the development and use of novel techniques. In contrast to many other scientific software projects, Dipy is not being developed by a single research group. Rather, it is an open project that encourages contributions from any scientist/developer through GitHub and open discussions on the project mailing list. Consequently, Dipy today has an international team of contributors, spanning seven different academic institutions in five countries and three continents, which is still growing.

Evaluation and statistical inference for human connectomes

Diffusion-weighted imaging coupled with tractography is currently the only method for in vivo mapping of human white-matter fascicles. Tractography takes diffusion measurements as input and produces the connectome, a large collection of white-matter fascicles, as output. We introduce a method to evaluate the evidence supporting connectomes. Linear fascicle evaluation (LiFE) takes any connectome as input and predicts diffusion measurements as output, using the difference between the measured and predicted diffusion signals to quantify the prediction error. We use the prediction error to evaluate the evidence that supports the properties of the connectome, to compare tractography algorithms and to test hypotheses about tracts and connections.

The vertical occipital fasciculus: A century of controversy resolved by in vivo measurements

Significance The vertical occipital fasciculus (VOF) is a major white-matter fascicle connecting dorsal and ventral visual cortex. Few vision scientists or cognitive neuroscientists are aware of the VOFs existence. The scarcity of papers on this important pathway stems from the contentious history surrounding its discovery by Wernicke in 1881. We review the conflict surrounding the classic, postmortem, VOF measurements, and we introduce modern, in vivo methods to precisely characterize the VOFs cortical terminations and unique tissue properties. The new VOF measurements provide insight into the communication between ventral stream regions involved in form perception and dorsal stream regions involved in eye movements and attention. The vertical occipital fasciculus (VOF) is the only major fiber bundle connecting dorsolateral and ventrolateral visual cortex. Only a handful of studies have examined the anatomy of the VOF or its role in cognition in the living human brain. Here, we trace the contentious history of the VOF, beginning with its original discovery in monkey by Wernicke (1881) and in human by Obersteiner (1888), to its disappearance from the literature, and recent reemergence a century later. We introduce an algorithm to identify the VOF in vivo using diffusion-weighted imaging and tractography, and show that the VOF can be found in every hemisphere (n = 74). Quantitative T1 measurements demonstrate that tissue properties, such as myelination, in the VOF differ from neighboring white-matter tracts. The terminations of the VOF are in consistent positions relative to cortical folding patterns in the dorsal and ventral visual streams. Recent findings demonstrate that these same anatomical locations also mark cytoarchitectonic and functional transitions in dorsal and ventral visual cortex. We conclude that the VOF is likely to serve a unique role in the communication of signals between regions on the ventral surface that are important for the perception of visual categories (e.g., words, faces, bodies, etc.) and regions on the dorsal surface involved in the control of eye movements, attention, and motion perception.

GLMdenoise: a fast, automated technique for denoising task-based fMRI data

In task-based functional magnetic resonance imaging (fMRI), researchers seek to measure fMRI signals related to a given task or condition. In many circumstances, measuring this signal of interest is limited by noise. In this study, we present GLMdenoise, a technique that improves signal-to-noise ratio (SNR) by entering noise regressors into a general linear model (GLM) analysis of fMRI data. The noise regressors are derived by conducting an initial model fit to determine voxels unrelated to the experimental paradigm, performing principal components analysis (PCA) on the time-series of these voxels, and using cross-validation to select the optimal number of principal components to use as noise regressors. Due to the use of data resampling, GLMdenoise requires and is best suited for datasets involving multiple runs (where conditions repeat across runs). We show that GLMdenoise consistently improves cross-validation accuracy of GLM estimates on a variety of event-related experimental datasets and is accompanied by substantial gains in SNR. To promote practical application of methods, we provide MATLAB code implementing GLMdenoise. Furthermore, to help compare GLMdenoise to other denoising methods, we present the Denoise Benchmark (DNB), a public database and architecture for evaluating denoising methods. The DNB consists of the datasets described in this paper, a code framework that enables automatic evaluation of a denoising method, and implementations of several denoising methods, including GLMdenoise, the use of motion parameters as noise regressors, ICA-based denoising, and RETROICOR/RVHRCOR. Using the DNB, we find that GLMdenoise performs best out of all of the denoising methods we tested.

Interactions of cognitive and auditory abilities in congenitally blind individuals

Congenitally blind individuals have been found to show superior performance in perceptual and memory tasks. In the present study, we asked whether superior stimulus encoding could account for performance in memory tasks. We characterized the performance of a group of congenitally blind individuals on a series of auditory, memory and executive cognitive tasks and compared their performance to that of sighted controls matched for age, education and musical training. As expected, we found superior verbal spans among congenitally blind individuals. Moreover, we found superior speech perception, measured by resilience to noise, and superior auditory frequency discrimination. However, when memory span was measured under conditions of equivalent speech perception, by adjusting the signal to noise ratio for each individual to the same level of perceptual difficulty (80% correct), the advantage in memory span was completely eliminated. Moreover, blind individuals did not possess any advantage in cognitive executive functions, such as manipulation of items in memory and math abilities. We propose that the short-term memory advantage of blind individuals results from better stimulus encoding, rather than from superiority at subsequent processing stages.

A Major Human White Matter Pathway Between Dorsal and Ventral Visual Cortex

Human visual cortex comprises many visual field maps organized into clusters. A standard organization separates visual maps into 2 distinct clusters within ventral and dorsal cortex. We combined fMRI, diffusion MRI, and fiber tractography to identify a major white matter pathway, the vertical occipital fasciculus (VOF), connecting maps within the dorsal and ventral visual cortex. We use a model-based method to assess the statistical evidence supporting several aspects of the VOF wiring pattern. There is strong evidence supporting the hypothesis that dorsal and ventral visual maps communicate through the VOF. The cortical projection zones of the VOF suggest that human ventral (hV4/VO-1) and dorsal (V3A/B) maps exchange substantial information. The VOF appears to be crucial for transmitting signals between regions that encode object properties including form, identity, and color and regions that map spatial information.

Cholinergic Enhancement Augments Magnitude and Specificity of Visual Perceptual Learning in Healthy Humans

Learning through experience underlies the ability to adapt to novel tasks and unfamiliar environments. However, learning must be regulated so that relevant aspects of the environment are selectively encoded. Acetylcholine (ACh) has been suggested to regulate learning by enhancing the responses of sensory cortical neurons to behaviorally relevant stimuli. In this study, we increased synaptic levels of ACh in the brains of healthy human subjects with the cholinesterase inhibitor donepezil (trade name: Aricept) and measured the effects of this cholinergic enhancement on visual perceptual learning. Each subject completed two 5 day courses of training on a motion direction discrimination task, once while ingesting 5 mg of donepezil before every training session and once while placebo was administered. We found that cholinergic enhancement augmented perceptual learning for stimuli having the same direction of motion and visual field location used during training. In addition, perceptual learning with donepezil was more selective to the trained direction of motion and visual field location. These results, combined with previous studies demonstrating an increase in neuronal selectivity following cholinergic enhancement, suggest a possible mechanism by which ACh augments neural plasticity by directing activity to populations of neurons that encode behaviorally relevant stimulus features.

Cholinergic enhancement increases the effects of voluntary attention but does not affect involuntary attention.

Voluntary visual spatial attention can be allocated in a goal-oriented manner to locations containing behaviorally relevant information. In contrast, involuntary attention is automatically captured by salient events. Allocation of attention is known to be modulated by release of the neurotransmitter acetylcholine (ACh) in cerebral cortex. We used an anti-predictive spatial cueing task to assess the effects of pharmacological enhancement of cholinergic transmission on behavioral measures of voluntary and involuntary attention in healthy human participants. Each trial began with the presentation of a cue in a peripheral location. In 80% of the trials, a target then appeared in a location opposite the cue. In the remaining 20% of trials, the target appeared in the cue location. For trials with short stimulus onset asynchrony (SOA) between cue and target, involuntary capture of attention resulted in shorter reaction times (RTs) to targets presented at the cue location. For long SOA trials, allocation of voluntary attention resulted in the opposite pattern: RTs were shorter when the target appeared in the expected (opposite) location. Each subject participated in two sessions: one in which the cholinesterase inhibitor donepezil was administered to increase synaptic ACh levels and one in which placebo was administered. Donepezil selectively improved performance (reduced RT) for long SOA trials in which targets appeared in the expected location. Thus, cholinergic enhancement augments the benefits of voluntary attention but does not affect involuntary attention, suggesting that they rely on different neurochemical mechanisms.

Human blindsight is mediated by an intact geniculo-extrastriate pathway.

Although damage to the primary visual cortex (V1) causes hemianopia, many patients retain some residual vision; known as blindsight. We show that blindsight may be facilitated by an intact white-matter pathway between the lateral geniculate nucleus and motion area hMT+. Visual psychophysics, diffusion-weighted magnetic resonance imaging and fibre tractography were applied in 17 patients with V1 damage acquired during adulthood and 9 age-matched controls. Individuals with V1 damage were subdivided into blindsight positive (preserved residual vision) and negative (no residual vision) according to psychophysical performance. All blindsight positive individuals showed intact geniculo-hMT+ pathways, while this pathway was significantly impaired or not measurable in blindsight negative individuals. Two white matter pathways previously implicated in blindsight: (i) superior colliculus to hMT+ and (ii) between hMT+ in each hemisphere were not consistently present in blindsight positive cases. Understanding the visual pathways crucial for residual vision may direct future rehabilitation strategies for hemianopia patients. DOI: http://dx.doi.org/10.7554/eLife.08935.001

A Two-Stage Cascade Model of BOLD Responses in Human Visual Cortex

Visual neuroscientists have discovered fundamental properties of neural representation through careful analysis of responses to controlled stimuli. Typically, different properties are studied and modeled separately. To integrate our knowledge, it is necessary to build general models that begin with an input image and predict responses to a wide range of stimuli. In this study, we develop a model that accepts an arbitrary band-pass grayscale image as input and predicts blood oxygenation level dependent (BOLD) responses in early visual cortex as output. The model has a cascade architecture, consisting of two stages of linear and nonlinear operations. The first stage involves well-established computations—local oriented filters and divisive normalization—whereas the second stage involves novel computations—compressive spatial summation (a form of normalization) and a variance-like nonlinearity that generates selectivity for second-order contrast. The parameters of the model, which are estimated from BOLD data, vary systematically across visual field maps: compared to primary visual cortex, extrastriate maps generally have larger receptive field size, stronger levels of normalization, and increased selectivity for second-order contrast. Our results provide insight into how stimuli are encoded and transformed in successive stages of visual processing.