Identification of band-pass filtered letters and faces by human and ideal observers

Contrast sensitivity for letter optotypes vs. gratings under conditions biased toward parvocellular and magnocellular pathways

This study examined the extent to which letter optotypes and grating stimuli provide equivalent measures of contrast sensitivity under conditions designed to favor the magnocellular (MC) and parvocellular (PC) pathways. The contrast sensitivity functions (CSFs) of three visually normal observers were measured for Sloan letters and Gabor patches, using steady- and pulsed-pedestal paradigms to bias processing toward MC and PC pathways, respectively. CSFs for Gabor patches were low-pass for the steady-pedestal paradigm and band-pass for the pulsed-pedestal paradigm, in agreement with previous reports. However, CSFs for letters were low-pass for both testing paradigms. CSFs for letters restricted in frequency content by spatial filtering were equivalent to those for Gabor patches for both testing paradigms. Results indicate that conventional letter optotypes can provide a misleading measure of contrast sensitivity, especially under conditions emphasizing the PC pathway. The use of spatially band-pass filtered letters can provide a more appropriate evaluation of spatial contrast sensitivity while maintaining some of the potential advantages of letters.

Quantifying facial expression recognition across viewing conditions

Facial expressions are key to social interactions and to assessment of potential danger in various situations. Therefore, our brains must be able to recognize facial expressions when they are transformed in biologically plausible ways. We used synthetic happy, sad, angry and fearful faces to determine the amount of geometric change required to recognize these emotions during brief presentations. Five-alternative forced choice conditions involving central viewing, peripheral viewing and inversion were used to study recognition among the four emotions. Two-alternative forced choice was used to study affect discrimination when spatial frequency information in the stimulus was modified. The results show an emotion and task-dependent pattern of detection. Facial expressions presented with low peak frequencies are much harder to discriminate from neutral than faces defined by either mid or high peak frequencies. Peripheral presentation of faces also makes recognition much more difficult, except for happy faces. Differences between fearful detection and recognition tasks are probably due to common confusions with sadness when recognizing fear from among other emotions. These findings further support the idea that these emotions are processed separately from each other.

Noise masking reveals channels for second-order letters

We investigate the channels underlying identification of second-order letters using a critical-band masking paradigm. We find that observers use a single 1-1.5 octave-wide channel for this task. This channel's best spatial frequency (c/letter) did not change across different noise conditions (indicating the inability of observers to switch channels to improve signal-to-noise ratio) or across different letter sizes (indicating scale invariance), for a fixed carrier frequency (c/letter). However, the channel's best spatial frequency does change with stimulus carrier frequency (both in c/letter); one is proportional to the other. Following Majaj et al. (Majaj, N. J., Pelli, D. G., Kurshan, P., & Palomares, M. (2002). The role of spatial frequency channels in letter identification. Vision Research, 42, 1165-1184), we define "stroke frequency" as the line frequency (strokes/deg) in the luminance image. That is, for luminance-defined letters, stroke frequency is the number of lines (strokes) across each letter divided by letter width. For second-order letters, letter texture stroke frequency is the number of carrier cycles (luminance lines) within the letter ink area divided by the letter width. Unlike the nonlinear dependence found for first-order letters (implying scale-dependent processing), for second-order letters the channel frequency is half the letter texture stroke frequency (suggesting scale-invariant processing).

Faces are "spatial"--holistic face perception is supported by low spatial frequencies

Faces are perceived holistically, a phenomenon best illustrated when the processing of a face feature is affected by the other features. Here, the authors tested the hypothesis that the holistic perception of a face mainly relies on its low spatial frequencies. Holistic face perception was tested in two classical paradigms: the whole-part advantage (Experiment 1) and the composite face effect (Experiments 2-4). Holistic effects were equally large or larger for low-pass filtered faces as compared to full-spectrum faces and significantly larger than for high-pass filtered faces. The disproportionate composite effect found for low-pass filtered faces was not observed when holistic perception was disrupted by inversion (Experiment 3). Experiment 4 showed that the composite face effect was enhanced only for low spatial frequencies, but not for intermediate spatial frequencies known be critical for face recognition. These findings indicate that holistic face perception is largely supported by low spatial frequencies. They also suggest that holistic processing precedes the analysis of local features during face perception.

Individual differences in FFA activity suggest independent processing at different spatial scales

The brain processes images at different spatial scales, but it is unclear how far into the visual stream different scales remain segregated. Using functional magnetic resonance imaging, we found evidence that BOLD activity in the fusiform face area (FFA) reflects computations based on separate spatial frequency inputs. When subjects perform different tasks (attend location vs. identity; attend whole vs. parts) or the same task with different stimuli (upright or inverted) with high- and low-pass images of cars and faces, individual differences in the FFA in one condition are correlated with those in the other condition. However, FFA activity in response to low-pass stimuli is independent of its response to high-pass stimuli. These results suggest that spatial scales are not integrated before the FFA and that processing in this area could support the flexible use of different sources of information present in broad-pass images.

Stimulus selectivity of figural aftereffects for faces

Viewing a distorted face induces large aftereffects in the appearance of an undistorted face. The authors examined the processes underlying this adaptation by comparing how selective the aftereffects are for different dimensions of the images including size, spatial frequency content, contrast, and color. Face aftereffects had weaker selectivity for changes in the size, contrast, or color of the images and stronger selectivity for changes in contrast polarity or spatial frequency. This pattern could arise if the adaptation is contingent on the perceived similarity of the stimuli as faces. Consistent with this, changing contrast polarity or spatial frequency had larger effects on the perceived identity of a face, and aftereffects were also selective for different individual faces. These results suggest that part of the sensitivity changes underlying the adaptation may arise at visual levels closely associated with the representation of faces.

Generic and customized digital image enhancement filters for the visually impaired

This study compares the effectiveness of various image enhancement filters for improving the perceived visibility of coloured digital natural images for people with visual impairment. Generic filters were compared with Peli's adaptive enhancement and adaptive thresholding and custom-devised filters based on each subject's contrast sensitivity loss. Subjects with low vision made within filter rankings followed by between filter ratings. In general, subjects preferred filters with lower gains. Unsharp masking resulted in a significant increase in perceived visibility for some image types (p < or = 0.05) while Peli's adaptive enhancement, edge enhancement and histogram equalization resulted in borderline improvements. Adaptive thresholding and the custom devised filter did not result in overall improvements in perceived visibility.

fMRI evidence for the neural representation of faces

fMRI (functional magnetic resonance imaging) studies on humans have shown a cortical area, the fusiform face area, that is specialized for face processing. An important question is how faces are represented within this area. This study provides direct evidence for a representation in which individual faces are encoded by their direction (facial identity) and distance (distinctiveness) from a prototypical (mean) face. When facial geometry (head shape, hair line, internal feature size and placement) was varied, the fMRI signal increased with increasing distance from the mean face. Furthermore, adaptation of the fMRI signal showed that the same neural population responds to faces falling along single identity axes within this space.

Configural masking of faces: Evidence for high-level interactions in face perception

The perception of a stimulus can be impaired when presented in the context of a masking pattern. To determine the timing and the nature of face processing, the effect of various masks on the discriminability of faces was investigated. Results reveal a strong configural effect: the magnitude of masking depends on the similarity between mask and target. Masking is absent for non-face masks (noise, houses), modest for scrambled and inverted faces and strongest for upright faces, even when they differ in size, gender or viewpoint from the targets. This suggests an extra-striate location for the masking (possibly FFA). Reduced but significant masking for isolated face parts (internal features or head shape) is consistent with holistic computations in face perception. The duration over which a face mask can impair face discrimination (130 ms) is markedly longer than previously assumed and is sufficient for iterative and feedback computations to be part of face processing.

Critical band masking in optic flow

Visual processing has been widely investigated with narrow band stimuli at low contrasts. We used a masking paradigm to examine how visual sensitivity under these conditions compares with the perception of the direction of heading in real scenes (i.e., with dynamic natural images at high contrasts). We first confirmed and extended previous studies showing biases in the amplitude distribution for spatial frequency, temporal frequency, speed and direction in dynamic natural movies. We then measured contrast thresholds for identification of the direction of motion for an observer traveling at various speeds. In spite of differences in contrast sensitivity and large non-uniformities in the amplitude content of the stimuli, contrast thresholds were relatively invariant of spatial frequency and completely invariant of temporal frequency, speed and direction. Our results suggest that visual processing normalises responses to supra-threshold structure at different spatial and temporal frequencies within natural stimuli and so equates their effective visibility.

The nature of synthetic face adaptation

Recent evidence demonstrates that adapting to a face will systematically bias the perception of faces that lie along the same identity trajectory in geometric face space but not faces that lie along different identity trajectories. We explored this configural aftereffect using synthetic face stimuli developed to measure face-specific processing. Adapting to synthetic "anti-faces" resulted in an identity-specific aftereffect that was characterized by a marked decrease in the slope of the psychometric functions. Adaptation transferred across different face sizes, but not different face viewpoints nor faces constructed about a non-mean face. Performance was captured by a model where responses were modulated through a divisive gain control and an additive constant reflecting a shift in the origin of perceived face space. Together, these results suggest that face adaptation reflects activity from mechanisms common to various processing stages along the visual pathway.

Face recognition is affected by similarity in spatial frequency range to a greater degree than within-category object recognition

Previous studies have suggested that face identification is more sensitive to variations in spatial frequency content than object recognition, but none have compared how sensitive the 2 processes are to variations in spatial frequency overlap (SFO). The authors tested face and object matching accuracy under varying SFO conditions. Their results showed that object recognition was more robust to SFO variations than face recognition and that the vulnerability of faces was not due to reliance on configural processing. They suggest that variations in sensitivity to SFO help explain the vulnerability of face recognition to changes in image format and the lack of a middle-frequency advantage in object recognition.

Spatial frequency requirements for audiovisual speech perception

Spatial frequency band-pass and low-pass filtered images of a talker were used in an audiovisual speech-in-noise task. Three experiments tested subjects' use of information contained in the different filter bands with center frequencies ranging from 2.7 to 44.1 cycles/face (c/face). Experiment 1 demonstrated that information from a broad range of spatial frequencies enhanced auditory intelligibility. The frequency bands differed in the degree of enhancement, with a peak being observed in a mid-range band (11-c/face center frequency). Experiment 2 showed that this pattern was not influenced by viewing distance and, thus, that the results are best interpreted in object spatial frequency, rather than in retinal coordinates. Experiment 3 showed that low-pass filtered images could produce a performance equivalent to that produced by unfiltered images. These experiments are consistent with the hypothesis that high spatial resolution information is not necessary for audiovisual speech perception and that a limited range of spatial frequency spectrum is sufficient.

Inversion Leads to Quantitative, Not Qualitative, Changes in Face Processing

Humans are remarkably adept at recognizing objects across a wide range of views. A notable exception to this general rule is that turning a face upside down makes it particularly difficult to recognize. This striking effect has prompted speculation that inversion qualitatively changes the way faces are processed. Researchers commonly assume that configural cues strongly influence the recognition of upright, but not inverted, faces. Indeed, the assumption is so well accepted that the inversion effect itself has been taken as a hallmark of qualitative processing differences. Here, we took a novel approach to understand the inversion effect. We used response classification to obtain a direct view of the perceptual strategies underlying face discrimination and to determine whether orientation effects can be explained by differential contributions of nonlinear processes. Inversion significantly impaired performance in our face discrimination task. However, surprisingly, observers utilized similar, local regions of faces for discrimination in both upright and inverted face conditions, and the relative contributions of nonlinear mechanisms to performance were similar across orientations. Our results suggest that upright and inverted face processing differ quantitatively, not qualitatively; information is extracted more efficiently from upright faces, perhaps as a by-product of orientation-dependent expertise.

No troubles with bubbles: a reply to Murray and Gold

Murray and Gold discuss two "shortcomings" of the Bubbles method [Vision Research 41 (2001) 2261]. The first one is theoretical: Bubbles would not fully characterize the LAM (Linear Amplifier Model) observer, whereas reverse correlation would. The second "shortcoming" is practical: the apertures that partly reveal information in a typical Bubbles experiment would induce atypical strategies in human observers, whereas the additive Gaussian white noise used by Murray and Gold (and others) in conjunction with reverse correlation would not. Here, we show that these claims are unfounded.

Troubles with bubbles

The bubbles method is a recently developed variant of reverse correlation methods that have been used in psychophysics and physiology. We show mathematically that for the broad and important class of noisy linear observers, the bubbles method recovers much less information about how observers process stimuli than reverse correlation does. We also show experimentally that the unusual type of noise used in the bubbles method can drastically change human observers' strategies in psychophysical tasks, which reduces the value of the information that is obtained from a bubbles experiment. We conclude that reverse correlation is generally preferable to the bubbles method in its present form, but we also give suggestions as to how the bubbles method could be modified to avoid the problems we discuss.

Expectancy effects on spatial frequency processing

We explored top-down modulation of spatial frequency (SF) processing. When auditory pre-cueing directed observers' attention to one of two 4-octaves (SF) apart plaid components observers tended to perceive the cued component, suggesting selective attention to the SF channel they expected to carry task relevant information. In agreement, pre-cueing had no effect with components often processed by the same channel (0.5-octaves apart). Further, effects of expectancy were greater than of uncertainty and were SF tuned. Combined our findings suggest top-down modulation of early, cortical, SF processing. We argue this could similarly explain the previously reported influences of categorisation on SF processing.

Utilisation of spatial frequency information in face search

In previous studies the utilisation of spatial frequency information in face perception has been investigated by using static recognition tasks. In this study we used a visual search task, which requires eye movements and fast identification of previously learned facial photographs. Using Fourier phase randomisation, spatial information was selectively removed without changing the amplitude spectrum of the image. Fourier phase was randomised within one-octave wide bands of nine different centre spatial frequencies (2-32 c/face width, 0.63-10.1 c/deg). In a control condition no randomisation was used. All stimuli had similar contrast. Search times and eye movements during the search were measured. The removal of spatial information by phase randomisation at medium spatial frequencies resulted in a considerable increase of search times. In the main experiment the maximum of the search times occurred between 8 and 11 c/face width. The number of eye fixations behaved similarly. In an additional experiment with a threefold viewing distance the search times increased and the maximum of the search times shifted slightly to lower object spatial frequencies (5.6-8 c/face width). This suggests that the band of spatial frequencies used in face search is not completely scale invariant. The results show that information most important to face search is located at a limited band of mid spatial frequencies. This is consistent with earlier studies, in which non-dynamical face recognition tasks and low-contrast stimuli have been used.

Effects on face recognition of spatial-frequency information contained in inspection and test stimuli

Researchers often assume a critical band of spatial frequencies is required for face recognition. Also, many studies have not measured the contrast required for recognition. On Day 1, observers viewed high-pass-filtered (HP), low-pass-filtered (LP), or unfiltered (UF) faces. On Day 2, they viewed a variety of faces, some of which were LP filtered, HP filtered, and UF. Observers adjusted contrast until they achieved both detection and recognition. Observers were most accurate and sensitive when filtered faces agreed in spatial-frequency content across days. Faces differing in spatial-frequency content were least well recognized. Unfiltered faces always fell between the 2 extremes. Observers generally used less contrast to recognize unfiltered than filtered faces. Correspondence of information between inspection and testing seemed more important than any particular range of frequencies.

Facial expression recognition in people with medicated and unmedicated Parkinson's disease

Recognition of facial expressions of emotion was investigated in people with medicated and unmedicated Parkinson's disease (PD) and matched controls (unmedicated PD, n=16; medicated PD, n=20; controls, n=40). Participants in the medicated group showed some visual impairment (impaired contrast sensitivity) and performed less well on perception of unfamiliar face identity, but did not show significant deficits in the perception of sex, gaze direction, or familiar identity from the face. For both Parkinson's disease groups, there was evidence of impaired recognition of facial expressions in comparison to controls. These deficits were more consistently noted in the unmedicated group, who were also found to perform worse than the medicated group at recognising disgust from prototypical facial expressions, and at recognising anger and disgust in computer-manipulated images. Although both Parkinson's disease groups showed impairments of facial expression recognition, the consistently worse recognition of disgust in the unmedicated group is consistent with the hypothesis from previous studies that brain regions modulated by dopaminergic neurons are involved in the recognition of disgust.

Scaling of letter size and contrast equalises perception across eccentricities and set sizes

Double E(2)N(2) scaling, i.e. magnifying size and contrast, allows modelling of the deterioration of face recognition performance with increasing eccentricity (E) and the size (N) of the set from which a target face has to be identified. E(2) and N(2) values represent the eccentricities and set sizes at which stimulus size and contrast must double in order to keep performance unchanged, whilst parameter K represents the multiplicative interaction between E and N. In the current study we investigated whether double E(2)N(2) scaling can model performance deterioration with increasing eccentricity and set size in letter perception too. Contrast sensitivity for letter perception was investigated as a function of letter size at N=1-8 and E=0 degrees -10 degrees. The superimposition of contrast sensitivity functions produced two scaling surfaces, one for letter size and another for contrast, which allowed modelling of the changes in letter perception with increasing E and N. With increasing eccentricity/set size the change of scale was much faster for contrast than letter size. Thus, in letter perception, contrast scaling was more important than spatial scaling. When compared with face perception, the change of spatial scale with increasing eccentricity was slower for letters whereas the change of contrast scale was similar for both. With increasing set size the changes of both spatial and contrast scales are faster for faces. In spatial scaling the interaction between eccentricity and set size was similar for letters and faces whereas in contrast scaling letters showed no interaction. Thus, letter perception was less affected by eccentricity and set size than face perception.

Learning to see faces and objects

Visual recognition of objects is an impressively difficult problem that biological systems solve effortlessly. We consider two aspects of this ability. First, is the recognition of all objects accomplished by either a single system or multiple, domain-specific systems? Behavioral, neuropsychological and neuroimaging data indicate that a single system is sufficient for the recognition of all objects at all levels. Second, how does such a system 'tune' itself to the constraints imposed by recognition at different levels of specificity? Evidence indicates that the task demands and learning that arise from different forms of feedback determine which computational routines are recruited automatically in object recognition.

Synthetic faces, face cubes, and the geometry of face space

To simplify the study of visual face processing, we introduce a novel class of synthetic face stimuli based upon 37 measurements (head shape, feature locations, etc.) extracted from individual face photographs in both frontal and 20 degrees side views. Synthetic faces are bandpass filtered optimally for face perception and include both line and edge information. Pilot experiments establish that subjects are extremely accurate in matching a synthetic face with the original grayscale photograph, even across views. To determine the perceptual metric of face space, we introduce face cubes in which the geometric differences between any faces in a four-dimensional face subspace can be precisely determined. Experiments on face discrimination using face cubes establish the metric of synthetic face space as locally Euclidean, with discrimination thresholds representing 4-6% total geometric variation (as a percent of mean head radius) between faces. Discrimination thresholds are lowest for face cubes constructed around the average face, thus indicating that the mean face for each gender represents a natural origin for face space. Finally, synthetic faces exhibit a pronounced inversion effect for 20 degrees side views and a characteristic "Thatcher effect" for inverted front views. Synthetic faces and face cubes thus provide a useful new quantitative approach to the study of face perception and face space.

Eccentricity bias as an organizing principle for human high-order object areas

We have recently proposed a center-periphery organization based on resolution needs, in which objects engaging in recognition processes requiring central-vision (e.g., face-related) are associated with center-biased representations, while objects requiring large-scale feature integration (e.g., buildings) are associated with periphery-biased representations. Here we tested this hypothesis by comparing the center-periphery organization with activations to five object categories: faces, buildings, tools, letter strings, and words. We found that faces, letter strings, and words were mapped preferentially within the center-biased representation. Faces showed a hemispheric lateralization opposite to that of letter strings and words. In contrast, buildings were mapped mainly to the periphery-biased representation, while tools activated both central and peripheral representations. The results are compatible with the notion that center-periphery organization allows the optimal allocation of cortical magnification to the specific requirements of various recognition processes.

The role of spatial frequency channels in letter identification

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 not free to choose which channels they use. We used critical-band masking to characterize the channels mediating identification of broadband signals: letters in a wide range of fonts (Sloan, Bookman, Künstler, Yung), alphabets (Roman and Chinese), and sizes (0.1-55 degrees ). We also tested sinewave and squarewave gratings. Masking always revealed a single channel, 1.6+/-0.7 octaves wide, with a center frequency that depends on letter size and alphabet. We define an alphabet's stroke frequency as the average number of lines crossed by a slice through a letter, divided by the letter width. For sharp-edged (i.e. broadband) signals, we find that stroke frequency completely determines channel frequency, independent of alphabet, font, and size. Moreover, even though observers have multiple channels, they always use the same channel for the same signals, even after hundreds of trials, regardless of whether the noise is low-pass, high-pass, or all-pass. This shows that observers identify letters through a single channel that is selected bottom-up, by the signal, not top-down by the observer. We thought shape would be processed similarly at all sizes. Bandlimited signals conform more to this expectation than do broadband signals. Here, we characterize processing by channel frequency. For sinewave gratings, as expected, channel frequency equals sinewave frequency f(channel)=f. For bandpass-filtered letters, channel frequency is proportional to center frequency f(channel) proportional, variantf(center) (log-log slope 1) when size is varied and the band (c/letter) is fixed, but channel frequency is less than proportional to center frequency f(channel) proportional, variantf(center)(2/3) (log-log slope 2/3) when the band is varied and size is fixed. Finally, our main result, for sharp-edged (i.e. broadband) letters and squarewaves, channel frequency depends solely on stroke frequency, f(channel)/10c/deg=(2/3), with a log-log slope of 2/3. Thus, large letters (and coarse squarewaves) are identified by their edges; small letters (and fine squarewaves) are identified by their gross strokes.

Identification of spatially filtered stimuli as function of the semantic category

The different weight of spatial frequency content in the identification of visual objects was investigated. Subjects were required to identify spatially filtered pictures of animals, vegetables and nonliving objects, displayed at 9 resolution levels of filtering following a coarse-to-fine order. Results showed that spatial frequency content differentially affected the three categories of stimuli. Data suggested a different involvement of low and high spatial frequency channels in visual processing of objects in relation to the semantic category.

Bubbles: a technique to reveal the use of information in recognition tasks

Everyday, people flexibly perform different categorizations of common faces, objects and scenes. Intuition and scattered evidence suggest that these categorizations require the use of different visual information from the input. However, there is no unifying method, based on the categorization performance of subjects, that can isolate the information used. To this end, we developed Bubbles, a general technique that can assign the credit of human categorization performance to specific visual information. To illustrate the technique, we applied Bubbles on three categorization tasks (gender, expressive or not and identity) on the same set of faces, with human and ideal observers to compare the features they used.

Size tuning in the absence of spatial frequency tuning in object recognition

How do we attend to objects at a variety of sizes as we view our visual world? Because of an advantage in identification of lowpass over highpass filtered patterns, as well as large over small images, a number of theorists have assumed that size-independent recognition is achieved by spatial frequency (SF) based coarse-to-fine tuning. We found that the advantage of large sizes or low SFs was lost when participants attempted to identify a target object (specified verbally) somewhere in the middle of a sequence of 40 images of objects, each shown for only 72 ms, as long as the target and distractors were the same size or spatial frequency (unfiltered or low or high bandpassed). When targets were of a different size or scale than the distractors, a marked advantage (pop out) was observed for large (unfiltered) and low SF targets against small (unfiltered) and high SF distractors, respectively, and a marked decrement for the complementary conditions. Importantly, this pattern of results for large and small images was unaffected by holding absolute or relative SF content constant over the different sizes and it could not be explained by simple luminance- or contrast-based pattern masking. These results suggest that size/scale tuning in object recognition was accomplished over the first several images (<576 ms) in the sequence and that the size tuning was implemented by a mechanism sensitive to spatial extent rather than to variations in spatial frequency.

Nonmonotonic noise tuning of BOLD fMRI signal to natural images in the visual cortex of the anesthetized monkey

BACKGROUND

The perceptual ability of humans and monkeys to identify objects in the presence of noise varies systematically and monotonically as a function of how much noise is introduced to the visual display. That is, it becomes more and more difficult to identify an object with increasing noise. Here we examine whether the blood oxygen level-dependent functional magnetic resonance imaging (BOLD fMRI) signal in anesthetized monkeys also shows such monotonic tuning. We employed parametric stimulus sets containing natural images and noise patterns matched for spatial frequency and intensity as well as intermediate images generated by interpolation between natural images and noise patterns. Anesthetized monkeys provide us with the unique opportunity to examine visual processing largely in the absence of top-down cognitive modulations and can thus provide an important baseline against which work with awake monkeys and humans can be compared.

RESULTS

We measured BOLD activity in occipital visual cortical areas as natural images and noise patterns, as well as intermediate interpolated patterns at three interpolation levels (25%, 50%, and 75%) were presented to anesthetized monkeys in a block paradigm. We observed reliable visual activity in occipital visual areas including V1, V2, V3, V3A, and V4 as well as the fundus and anterior bank of the superior temporal sulcus (STS). Natural images consistently elicited higher BOLD levels than noise patterns. For intermediate images, however, we did not observe monotonic tuning. Instead, we observed a characteristic V-shaped noise-tuning function in primary and extrastriate visual areas. BOLD signals initially decreased as noise was added to the stimulus but then increased again as the pure noise pattern was approached. We present a simple model based on the number of activated neurons and the strength of activation per neuron that can account for these results.

CONCLUSIONS

We show that, for our parametric stimulus set, BOLD activity varied nonmonotonically as a function of how much noise was added to the visual stimuli, unlike the perceptual ability of humans and monkeys to identify such stimuli. This raises important caveats for interpreting fMRI data and demonstrates the importance of assessing not only which neural populations are activated by contrasting conditions during an fMRI study, but also the strength of this activation. This becomes particularly important when using the BOLD signal to make inferences about the relationship between neural activity and behavior.

The effects of aging on visual memory: evidence for functional reorganization of cortical networks

Recent evidence suggests that the mature human brain is capable of substantial functional reorganization following injury. The fact that the brain retains a great deal of plasticity raises the possibility that cortical reorganization may occur during normal aging. We examined this issue by using positron emission tomography (PET) to measure the brain activity associated with short-term memory for simple visual attributes in young and old observers. A two-interval forced choice procedure was used to measure spatial frequency discrimination thresholds for sine wave gratings presented at different inter-stimulus intervals (ISI). Memory load was manipulated by varying the duration of the ISI and by presenting an irrelevant masking stimulus in the middle of the ISI. Old and young observers performed the experiment equally well. However, the neural systems correlated with good performance differed for the two age groups. The results support the hypothesis that the functional networks that underlie visual memory undergo reorganization during aging.

Identification of facial images in peripheral vision

Contrast sensitivity for face identification was measured as a function of image size to find out whether foveal and peripheral performance would become equivalent by magnification. Size scaling was not sufficient for this task, but when the data was scaled both in size and contrast dimensions, there was no significant eccentricity-dependent variation in the data, i.e. for equivalent performance both the size and contrast needed to increase in the periphery. By utilising spatial noise added to the images we found that in periphery information was utilised less efficiently and peripheral inferiority arose completely from lower efficiency, not from increased internal noise.

Signal but not noise changes with perceptual learning

Perceptual discrimination improves with practice. This 'perceptual learning' is often specific to the stimuli presented during training, indicating that practice may alter the response characteristics of cortical sensory neurons. Although much is known about how learning modifies cortical circuits, it remains unclear how these changes relate to behaviour. Different theories assume that practice improves discrimination by enhancing the signal, diminishing internal noise or both. Here, to distinguish among these alternatives, we fashioned sets of faces and textures whose signal strength could be varied, and we trained observers to identify these patterns embedded in noise. Performance increased by up to 400% across several sessions over several days. Comparisons of human performance to that of an ideal discriminator showed that learning increased the efficiency with which observers encoded task-relevant information. Observer response consistency, measured by a double-pass technique in which identical stimuli are shown twice in each experimental session, did not change during training, showing that learning had no effect on internal noise. These results indicate that perceptual learning may enhance signal strength, and provide important constraints for theories of learning.