bioRxiv Science⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/080580

Inferring exemplar discriminability in brain representations

Hamed Nili·Alexander Walther·Arjen Alink·Nikolaus Kriegeskorte
View original source ↗Export metadata ↓

Abstract

Representational distinctions within categories are important in all perceptual modalities and also in cognitive and motor representations. Recent pattern-information studies of brain activity have used condition-rich designs to sample the stimulus space more densely. To test whether brain response patterns discriminate among a set of stimuli (e.g. exemplars within a category) with good sensitivity, we can pool statistical evidence over all pairwise comparisons. A popular test statistic reflecting exemplar information is the exemplar discriminability index (EDI), which is defined as the average of the pattern dissimilarity estimates between different exemplars minus the average of the pattern dissimilarity estimates between repetitions of identical exemplars. The EDI is commonly tested with a t test (H0: population mean EDI = 0) across subjects (subject as random effect). However, it is unclear whether this approach is either valid or optimal. Here we describe a wide range of statistical tests of exemplar discriminability and assess the validity (specificity) and power (sensitivity) of each test. The tests include previously used and novel, parametric and nonparametric tests, which treat subject as a random or fixed effect, and are based on different dissimilarity measures, different test statistics, and different inference procedures. We use simulated and real data to determine which tests are valid and which are most sensitive. The popular across-subject t test of the EDI (typically using correlation distance as the pattern dissimilarity measure) requires the assumption that the EDI is 0-mean normal under H0, which is not strictly true. Reassuringly, our simulations suggest that the test controls the false-positives rate at the nominal level and is thus valid in practice. However, test statistics based on average Mahalanobis distances or average linear-discriminant t values (both accounting for the multivariate error covariance among responses) are substantially more powerful for both random- and fixed-effects inference. We suggest preferred procedures for safely and sensitively detecting subtle pattern differences between exemplars.

Source connections

Explore related subjects

NeurosciencePreprint

Keep this discovery

BibTeXRIS

Hamed Nili, Alexander Walther, Arjen Alink, Nikolaus Kriegeskorte. 2016-10-12. Inferring exemplar discriminability in brain representations. https://doi.org/10.1101/080580

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2016

Temporal and spatial localization of prediction-error signals in the visual brain

It has been suggested that the brain pre-empts changes in the visual environment through generating predictions, although real-time eletrophysiological evidence of prediction violations remains elusive. In a series of experiments we showed participants sequences of images that followed a predictable implied sequence or whose final image violated the implied sequence. Through careful design we were able to use the same final image transitions across predictable and unpredictable conditions, ensuring that any differences in neural responses were due only to preceding context and not to the images themselves. EEG and MEG recordings showed that early/mid-latency visual evoked potentials were robustly modulated by images that violated the implied sequence across a range of types of image change (expression deformations, rigid-rotations and visual field location). This modulation occurred irrespective of stimulus object category. Although the stimuli were static images, MEG source reconstruction of the early latency signal (N/M170) localised expectancy violation signals to brain areas associated with motion perception. Our findings suggest that the N/M170 can index mismatches between predicted and actual visual inputs in a system that predicts trajectories based on ongoing context. This has important implications for understanding the N/M170 and investigating how the brain represents context to generate perceptual predictions.

Neuroscience↗
bioRxiv2016

AN OSCILLATORY NETWORK MODEL OF HEAD DIRECTION, SPATIALLY PERIODIC CELLS AND PLACE CELLS USING LOCOMOTOR INPUTS

We propose a computational modeling approach that explains the formation of a range of spatial cells like head direction cells, grid cells, border cells and place cells which are believed to play a pivotal role in the spatial navigation of an animal. Most existing models insert special symmetry conditions in the models in order to obtain such symmetries in the outcome; our models do not require such symmetry assumptions. Our modeling approach is embodied in two models: a simple one (Model #1) and a more detailed version (Model #2). In Model #1, velocity input is presented to a layer of Head Direction cells, with no special topology requirements, the outputs of which are presented to a layer of Path Integration neurons. A variety of spatially periodic responses resembling grid cells, are obtained using the Principal Components of Path Integration layer. In Model #2, the input consists of the locomotor rhythms from the four legs of a virtual animal. These rhythms are integrated into the phases of a layer of oscillatory neurons, whose outputs drive a layer of Head Direction cells. The Head Direction cells in turn drive a layer of Path Integration neurons, which in turn project to two successive layers of Lateral Anti Hebbian Networks (LAHN). Cells in the first LAHN resemble grid cells (with both hexagonal and square gridness), and border cells. Cells in the second LAHN exhibit place cell behaviour and a new cell type known as corner cell. Both grid cells and place cells exhibit phase precession in 1D and 2D spaces. The models outline the neural hierarchy necessary to obtain the complete range of spatial cell responses found in the hippocampal system.

Neuroscience↗
bioRxiv2016

Association of polygenic risk for major psychiatric illness with subcortical volumes and white matter integrity in UK Biobank

Major depressive disorder (MDD), schizophrenia (SCZ) and bipolar disorder (BP) are common, disabling and heritable psychiatric diseases with a complex overlapping polygenic architecture. Individuals with these disorders, as well as their unaffected relatives, show widespread structural differences in corticostriatal and limbic networks. Structural variation in many of these brain regions is also heritable and polygenic but whether their genetic architecture overlaps with major psychiatric disorders is unknown. We sought to address this issue by examining the impact of polygenic risk of MDD, SCZ, and BP on subcortical brain volumes and white matter (WM) microstructure in a large single sample of neuroimaging data; the UK Biobank Imaging study. The first release of UK Biobank imaging data compromised participants with overlapping genetic data and subcortical volumes (N = 978) and WM measures (N = 816). Our, findings however, indicated no statistically significant associations between either subcortical volumes or WM microstructure, and polygenic risk for MDD, SCZ or BP. In the current study, we found little or no evidence for genetic overlap between major psychiatric disorders and structural brain measures. These findings suggest that subcortical brain volumes and WM microstructure may not be closely linked to the genetic mechanisms of major psychiatric disorders.

Neuroscience↗