Najib J Majaj

Najib J Majaj
New York University | NYU · Center for Neural Science (CNS)

Phd

About

73
Publications
8,605
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
3,399
Citations
Citations since 2017
24 Research Items
1782 Citations
2017201820192020202120222023050100150200250300350
2017201820192020202120222023050100150200250300350
2017201820192020202120222023050100150200250300350
2017201820192020202120222023050100150200250300350

Publications

Publications (73)
Preprint
Full-text available
Natural images contain information at multiple spatial scales. Although we understand how early visual mechanisms split multi-scale images into distinct spatial frequency channels, we do not know how the outputs of these channels are processed further by mid-level visual mechanisms. We have recently developed a naturalness discrimination task that...
Preprint
Full-text available
The core of everyday tasks like reading and driving is active object recognition. Attempts to model such tasks are currently stymied by the inability to incorporate time. People show a flexible tradeoff between speed and accuracy and this tradeoff is a crucial human skill. Deep neural networks have emerged as promising candidates for predicting pea...
Preprint
Full-text available
Crowding is the failure to recognize an object due to surrounding clutter. Its strength varies across the visual field and individuals. To characterize the statistics of crowding—ultimately to relate psychophysics of crowding to physiology—we measured radial crowding distance and acuity of 105 observers along the four cardinal meridians of the visu...
Preprint
Full-text available
Neural networks today often recognize objects as well as people do, and thus might serve as models of the human recognition process. However, most such networks provide their answer after a fixed computational effort, whereas human reaction time varies, e.g. from 0.2 to 10 s, depending on the properties of stimulus and task. To model the effect of...
Preprint
Scene text recognition models have advanced greatly in recent years. Inspired by human reading we characterize two important scene text recognition models by measuring their domains i.e. the range of stimulus images that they can read. The domain specifies the ability of readers to generalize to different word lengths, fonts, and amounts of occlusi...
Preprint
Full-text available
Deep convolutional artificial neural networks (ANNs) are the leading class of candidate models of the mechanisms of visual processing in the primate ventral stream. While initially inspired by brain anatomy, over the past years, these ANNs have evolved from a simple eight-layer architecture in AlexNet to extremely deep and branching architectures,...
Article
Full-text available
Many vision science studies employ machine learning, especially the version called "deep learning." Neuroscientists use machine learning to decode neural responses. Perception scientists try to understand how living organisms recognize objects. To them, deep neural networks offer benchmark accuracies for recognition of learned stimuli. Originally m...
Preprint
Full-text available
The internal representations of early deep artificial neural networks (ANNs) were found to be remarkably similar to the internal neural representations measured experimentally in the primate brain. Here we ask, as deep ANNs have continued to evolve, are they becoming more or less brain-like? ANNs that are most functionally similar to the brain will...
Preprint
Today many vision-science presentations employ machine learning, especially the version called “deep learning”. Many neuroscientists use machine learning to decode neural responses. Many perception scientists try to understand how living organisms recognize objects. To them, deep neural networks offer benchmark accuracies for recognition of learned...
Poster
Full-text available
Most studies of neuronal responses in amblyopia have focused on primary visual cortex (V1) in anesthetized animals. Typically these experiments reveal decreased binocularity, and reduced contrast sensitivity and visual resolution in cells driven by the amblyopic eye (AE) compared to the fellow eye (FE). These neuronal deficits are less marked than...
Article
In amblyopia, abnormal visual experience leads to an extreme form of eye dominance, in which vision through the nondominant eye is degraded. A key aspect of this disorder is perceptual suppression: the image seen by the stronger eye often dominates during binocular viewing, blocking the image of the weaker eye from reaching awareness. Interocular s...
Article
In amblyopia, a visual disorder caused by abnormal visual experience during development, the amblyopic eye loses visual sensitivity while the other (fellow) eye is largely unaffected. Binocular vision in amblyopes is often disrupted by interocular suppression. We used 96-electrode arrays to record neurons and neuronal groups in areas V1 and V2 of 6...
Article
Amblyopia is a developmental disorder leading to form vision deficits in the affected eye and interocular perceptual suppression. Monocular testing has revealed neuronal response patterns that can partly account for the form deficits, but few have studied amblyopic cortical responses under binocular viewing conditions. We wondered whether dichoptic...
Article
Full-text available
Spoken language is a central part of our everyday lives, but the precise roles that individual cortical regions play in the production of speech are often poorly understood. To address this issue, we focally lowered the temperature of distinct cortical regions in awake neurosurgical patients, and we relate this perturbation to changes in produced s...
Article
Extensive research has revealed that the ventral visual stream hierarchically builds a robust representation for supporting visual object categorization tasks. We systematically explored the ability of multiple ventral visual areas to support a variety of 'category-orthogonal' object properties such as position, size and pose. For complex naturalis...
Article
To go beyond qualitative models of the biological substrate of object recognition, we ask: can a single ventral stream neuronal linking hypothesis quantitatively account for core object recognition performance over a broad range of tasks? We measured human performance in 64 object recognition tests using thousands of challenging images that explore...
Article
Full-text available
Amblyopia is a developmental disorder resulting in poor vision in one eye. The mechanism by which input to the affected eye is prevented from reaching the level of awareness remains poorly understood. We recorded simultaneously from large populations of neurons in the supragranular layers of areas V1 and V2 in 6 macaques that were made amblyopic by...
Article
Amblyopia is a developmental disorder causing form vision deficits in the affected eye. Human psychophysics has shown that the representations of both luminance-defined and contrast-defined "2nd-order" forms are affected. Neurons in V1 and V2 of normal macaques typically show surround suppression (SS): the response to a grating is attenuated when t...
Article
Amblyopia is a developmental disorder in which spatial vision is impaired in one eye. The physiological basis for amblyopia is unknown, but previous work has shown that the amblyopic eye often influences fewer cortical neurons than the non-amblyopic eye, and that these neurons prefer lower spatial frequencies on average. Earlier studies relied on s...
Article
Amblyopia is a visual disorder associated with disruptions of conjugate binocular vision in early life, which cause a loss in visual performance usually in one eye. Previous studies explored the neurophysiological basis of the amblyopic deficit using single-unit recordings, but the results incompletely explain the perceptual deficits. Advances in r...
Article
Full-text available
Many experiments in neuroscience require or would benefit tremendously from a wireless neural recording system. However, commercially available wireless systems are expensive, have moderate to high noise and are often not customizable. Academic wireless systems present impressive capabilities [1]-[4], but are not available for other labs to use. To...
Article
Full-text available
The primate visual system achieves remarkable visual object recognition performance even in brief presentations and under changes to object exemplar, geometric transformations, and background variation (a.k.a. core visual object recognition). This remarkable performance is mediated by the representation formed in inferior temporal (IT) cortex. In p...
Article
Full-text available
A key requirement for the development of effective learning representations is their evaluation and comparison to representations we know to be effective. In natural sensory domains, the community has viewed the brain as a source of inspiration and as an implicit benchmark for success. However, it has not been possible to directly test representati...
Article
Schwartz et al. (ARVO '98) found that observers asked to identify facial expressions use only the mouth area of the face. Paul Ekman objected that her faces were not expressing emotions, and predicted that for emotive faces, observers would use both the mouth AND the eyes. To test his prediction, we measured thresholds for identifying emotions in t...
Article
Simoncelli & Heeger (1998) proposed a model of responses in MT that incorporates two stages corresponding to cortical areas V1 and MT; each stage includes a divisive contrast gain control or normalization. The model accounts well for many properties of MT neurons, but direct physiological evidence for a second stage of contrast gain control is lack...
Article
We have investigated the adaptation properties of single neurons in area MT of anesthetized macaque monkeys using drifting coherent dot patterns. We compared the direction tuning properties of neurons before and after an adapting stimulus (40s) whose direction was either in the preferred, null, or near-preferred direction (30 degrees from optimal)....
Article
Watson and Robson (1981) showed that observers can discriminate large frequency differences at detection threshold, but require more contrast to discriminate fine differences. They suggested that the extra contrast allows the observer to use two channels. However, critical band masking of identification of letters, objects, faces, and gratings alwa...
Article
Olzak and Thomas posited "cigar" channels, which integrate across a wide range of frequency and a narrow range of orientation, and "donut" channels, which integrate across all orientations and a narrow range of frequency. Majaj et al. (Vision Research, in press) used critical band masking to measure the radial frequency tuning of the channel that o...
Article
Introduction: Features of faces and words (e.g. eyes and letters) are more easily identified when they are presented in their familiar context as opposed to being presented alone or in an unfamiliar arrangement . In the fovea, this is known as the word and face superiority effect, and is evidence that words and faces are special (Reicher, 1969 Jour...
Article
At low speeds and contrasts, bar textures moving obliquely to their orientation change their apparent direction. Initially they seem to move at right angles to the bar orientation, then over ∼200 ms they change and appear to lock to the true direction of motion defined by the bars' ends ("terminator motion"). Responses in macaque MT neurons behave...
Article
We grew up thinking that all our channels are always available, waiting to be used, like the strings in a piano. Majaj et al. (2002 & this VSS) show that observers identifying letters or reading use only one spatial frequency channel, determined by the stroke frequency of the letters. We show that this is also true for faces and line drawings of co...
Article
Pelli et al. (2002) show that efficiency for identifying letters is inversely proportional to their perimetric complexity (perimeter squared over "ink" area). Here we report a much larger effect of complexity in the periphery. Efficiency for simple letters is similar in fovea and periphery, but efficiency for complex letters is a factor of five wor...
Article
Letter identification is mediated by just one spatial frequency channel (Solomon & Pelli, 1994). But what about reading? We wondered whether larger features, e.g. words, at lower spatial frequencies are used when reading text as opposed to just identifying letters. We characterized the channel for reading by measuring reading rate as a function of...
Article
Neurons in area MT can integrate the motion of 1D image elements to signal the 2D motion of complex patterns. This pattern direction selectivity is typically studied using plaids composed of two identical, high-contrast gratings. We wondered how pattern motion computation would be affected when we reduced the contrast of one of the plaid's componen...
Article
The old discovery that readers make several fixations per second, rather than a continuous sweep across the text, suggests that reading might be limited by the number of letters that can be acquired in one fixation. That span has been measured in various ways, but remains unexplained. Here we prove that the “visual span” is simply the number of cha...
Article
Full-text available
Analysis of the movement of a complex visual stimulus is expressed in the responses of pattern-direction-selective neurons in area MT, which depend in turn on directionally selective inputs from area V1. How do MT neurons integrate their inputs? Pattern selectivity in MT breaks down when the gratings comprising a moving plaid are presented to non-o...
Chapter
Early in visual processing neurons with small receptive fields can only signal the component of motion perpendicular to the orientation of the contour that passes through them (the “aperture problem”). A moving visual pattern with differently oriented contours can thus elicit neuronal responses that convey conflicting motion cues. To recover the tr...
Article
A favorite question in cognitive psychology about the computation underlying object recognition is whether we identify an object as a whole or by its parts. This deceptively simple question has been surprisingly hard to answer. Here we provide a quick diagnostic test, applicable to any object identification task that can be done in the periphery at...
Article
Full-text available
Magnocellular (M-), but not parvocellular (P-), neurons of the macaque lateral geniculate nucleus (LGN) differ distinctively in their responses to counterphase-modulated and drifting gratings. Relative to stimulation with drifting gratings, counterphase modulation reduces the responses of M- cells in a band around 25 Hz, producing a "notch" in the...
Article
Full-text available
The Gestalt psychologists reported a set of laws describing how vision groups elements to recognize objects. The Gestalt laws “prescribe for us what we are to recognize ‘as one thing’” (Köhler, 192024. Köhler , W. 1920. Die physischen Gestalten in Ruhe und im stationären Zustand, Erlangen, , Germany: Verlag der Philosophischen Akademie. [The physi...
Article
Full-text available
We characterize a hitherto undocumented type of neuron present in the regions bordering the principal layers of the macaque lateral geniculate nucleus. Neurons of this type were distinguished by a high and unusually regular maintained discharge that was suppressed by spatiotemporal modulation of luminance or chromaticity within the receptive field....
Article
Full-text available
Bouma's law of crowding predicts an uncrowded central window through which we can read and a crowded periphery through which we cannot. The old discovery that readers make several fixations per second, rather than a continuous sweep across the text, suggests that reading is limited by the number of letters that can be acquired in one fixation, with...
Article
Full-text available
Direction-selective neurons in primary visual cortex have small receptive fields that encode the motions of local features. These motions often differ from the motion of the object to which they belong and must therefore be integrated elsewhere. A candidate site for this integration is visual cortical area MT (V5), in which cells with large recepti...
Article
Psychophysical studies assign special status to three post-receptoral mechanisms of color vision: two color-opponent ones capturing red-green and yellow-blue variations, and a non-opponent one capturing achromatic variations. The distinctiveness of the cardinal mechanisms is most clearly revealed by habituation to chromatic modulation, but their lo...
Article
Most neurons in macaque area MT are selective for the direction of stimulus motion. By comparing direction selectivity for gratings and plaids, we classified MT neurons as pattern direction selective (PDS) or component direction selective (CDS). We compared the time course of responses in CDS and PDS neurons in opiate-anesthetized macaques, using a...
Article
Full-text available
Do we identify an object as a whole or by its parts? This simple question has been surprisingly hard to answer. It has been suggested that faces are recognized as wholes and words are recognized by parts. Here we answer the question by applying a test for crowding. In crowding, a target is harder to identify in the presence of nearby flankers. Prev...
Article
Full-text available
A letter in the peripheral visual field is much harder to identify in the presence of nearby letters. This is called "crowding". In general, masking is a procedure: introducing any "mask" pattern that affects discriminability of the signal. Crowding conforms to the masking paradigm, but the crowding effect is unlike ordinary masking. Here we charac...
Article
How we see is today explained by physical optics and retinal transduction, followed by feature detection, in the cortex, by a bank of parallel independent spatial-frequency-selective channels. It is assumed that the observer uses whichever channels are best for the task at hand. Our current results demand a revision of this framework: Observers are...
Article
It is well known that observers classifying signals varying along a single sensory dimension can only distinguish 7±2 categories (Miller, 1956). We wondered whether eye movements could escape this apparently cognitive bottleneck. In a dark room, the observer fixated a point of light, which was suddenly displaced horizontally by a random offset, and...
Article
Using critical-band masking, we characterized the channels mediating identification of letters in several fonts and alphabets (Bookman, Künstler, Sloan, and Chinese) over a wide range of size (0.07 to 10 deg). Line frequency is the average number of lines crossed by a rule through a letter, divided by the letter size. We expected channel frequency...
Article
For sinusoidal grating signals on a noise-free background, it is well known that the contrast sensitivity function (csf) has a log-log slope of +1 at low spatial frequencies Equivalent input noise measurements show that this slope is account 3d for by high intrinsic visual noise at low spatial frequencies (Raghavan et al. ARVO'95). However, it is a...

Network

Cited By