Early interactions between neuronal adaptation and voluntary control determine perceptual choices in bistable vision.

At the onset of bistable stimuli, the brain needs to choose which of the competing perceptual interpretations will first reach awareness. Stimulus manipulations and cognitive control both influence this choice process, but the underlying mechanisms and interactions remain poorly understood. Using intermittent presentation of bistable visual stimuli, we demonstrate that short interruptions cause perceptual reversals upon the next presentation, whereas longer interstimulus intervals stabilize the percept. Top-down voluntary control biases this process but does not override the timing dependencies. Extending a recently introduced low-level neural model, we demonstrate that percept-choice dynamics in bistable vision can be fully understood with interactions in early neural processing stages. Our model includes adaptive neural processing preceding a rivalry resolution stage with cross-inhibition, adaptation, and an interaction of the adaptation levels with a neural baseline. Most importantly, our findings suggest that top-down attentional control over bistable stimuli interacts with low-level mechanisms at early levels of sensory processing before perceptual conflicts are resolved and perceptual choices about bistable stimuli are made.

Percept-choice sequences driven by interrupted ambiguous stimuli: a low-level neural model.

Existing neural explanations of spontaneous percept switching under steady viewing of an ambiguous stimulus do not fit the fact that stimulus interruptions cause the same percept to reappear across many ON/OFF cycles. We present a simple neural model that explains the observed behavior and predicts several more complicated percept sequences, without invoking any "high-level" decision making or memory. Percept choice at stimulus onset, which differs fundamentally from standard percept switching, depends crucially on a hitherto neglected interaction between local "shunting" adaptation and a near-threshold neural baseline. Stimulus ON/OFF timing then controls the generation of repeating, alternating, or more complex choice sequences. Our model also explains "priming" versus "habituation" effects on percept choice, reinterprets recent neurophysiological data, and predicts the emergence of hysteresis at the level of percept sequences, with occasional noise-induced sequence "hopping."

Direct extraction of curvature-based metric shape from stereo by view-modulated receptive fields.

Any computation of metric surface structure from horizontal disparities depends on the viewing geometry, and analysing this dependence allows us to narrow down the choice of viable schemes. For example, all depth-based or slant-based schemes (i.e. nearly all existing models) are found to be unrealistically sensitive to natural errors in vergence. Curvature-based schemes avoid these problems and require only moderate, more robust view-dependent corrections to yield local object shape, without any depth coding. This fits the fact that humans are strikingly insensitive to global depth but accurate in discriminating surface curvature. The latter also excludes coding only affine structure. In view of new adaptation results, our goal becomes to directly extract retinotopic fields of metric surface curvatures (i.e. avoiding intermediate disparity curvature). To find a robust neural realisation, we combine new exact analysis with basic neural and psychophysical constraints. Systematic, step-by-step 'design' leads to neural operators which employ a novel family of 'dynamic' receptive fields (RFs), tuned to specific (bi-)local disparity structure. The required RF family is dictated by the non-Euclidean geometry that we identify as inherent in cyclopean vision. The dynamic RF-subfield patterns are controlled via gain modulation by binocular vergence and version, and parameterised by a cell-specific tuning to slant. Our full characterisation of the neural operators invites a range of new neurophysiological tests. Regarding shape perception, the model inverts widely accepted interpretations: It predicts the various types of errors that have often been mistaken for evidence against metric shape extraction.

Attentional control over either of the two competing percepts of ambiguous stimuli revealed by a two-parameter analysis: means do not make the difference.

We studied distributions of perceptual rivalry reversals, as defined by the two fitted parameters of the Gamma distribution. We did so for a variety of bi-stable stimuli and voluntary control exertion tasks. Subjects' distributions differed from one another for a particular stimulus and control task in a systematic way that reflects a constraint on the describing parameters. We found a variety of two-parameter effects, the most important one being that distributions of subjects differ from one another in the same systematic way across different stimuli and control tasks (i.e., a fast switcher remains fast across all conditions in a parameter-specified way). The cardinal component of subject-dependent variation was not the conventionally used mean reversal rate, but a component that was oriented-for all stimuli and tasks-roughly perpendicular to the mean rate. For the Necker cube, we performed additional experiments employing specific variations in control exertion, suggesting that subjects have to a considerable extent independent control over the reversal rate of either of the two competing percepts.

Designing lymphocyte functional structure for optimal signal detection: voilà, T cells.

One basic task of immune systems is to detect signals from unknown "intruders" amidst a noisy background of harmless signals. To clarify the functional importance of many observed lymphocyte properties, I ask: What properties would a cell have if one designed it according to the theory of optimal detection, with minimal regard for biological constraints? Sparse and reasonable assumptions about the statistics of available signals prove sufficient for deriving many features of the optimal functional structure, in an incremental and modular design. The use of one common formalism guarantees that all parts of the design collaborate to solve the detection task. Detection performance is computed at several stages of the design. Comparison between design variants reveals e.g. the importance of controlling the signal integration time. This predicts that an appropriate control mechanism should exist. Comparing the design to reality, I find a striking similarity with many features of T cells. For example, the formalism dictates clonal specificity, serial receptor triggering, (grades of) anergy, negative and positive selection, co-stimulation, high-zone tolerance, and clonal production of cytokines. Serious mismatches should be found if T cells were hindered by mechanistic constraints or vestiges of their (co-)evolutionary history, but I have not found clear examples. By contrast, fundamental mismatches abound when comparing the design to immune systems of e.g. invertebrates. The wide-ranging differences seem to hinge on the (in)ability to generate a large diversity of receptors.

Normal telomere lengths in naive and memory CD4+ T cells in HIV type 1 infection: a mathematical interpretation.

To study CD4+ T cell productivity during HIV-1 infection, CD4+ T cell telomere lengths were measured. Cross-sectional and longitudinal analysis of HIV-1-infected individuals with CD4+ T cells counts >300 cells/mm3 showed normal average telomeric restriction fragment (TRF) length and normal shortening rates of CD45RA+ naive and CD45RO+ memory CD4+ T cells. These TRF data were interpreted in terms of CD4+ T cell production by means of a mathematical model. This model resolves previous criticisms arguing that the normal TRF length of CD4+ T cells in HIV-1 clinical latency is due to the killing of dividing CD4+ T cells by the virus. Only an increased priming rate of naive CD4+ T cells to become memory cells may elongate the average TRF length of memory CD4+ T cells, and may therefore mask the shortening effect of increased turnover in the CD4+ memory T cell compartment. The data are more compatible with the notion that during HIV-1 clinical latency the turnover of CD4+ T cells is not markedly increased, however, and that HIV-related interference with renewal from progenitors plays a role in CD4+ T cell depletion. In such a "limited renewal" scenario disease progression is no longer a consequence of markedly increased CD4+ T cell production.

How specific should immunological memory be?

Protection against infection hinges on a close interplay between the innate immune system and the adaptive immune system. Depending on the type and context of a pathogen, the innate system instructs the adaptive immune system to induce an appropriate immune response. Here, we hypothesize that the adaptive immune system stores these instructions by changing from a naive to an appropriate memory phenotype. In a secondary immune reaction, memory lymphocytes adhere to their instructed phenotype. Because cross-reactions with unrelated Ags can be detrimental, such a qualitative form of memory requires a sufficient degree of specificity of the adaptive immune system. For example, lymphocytes instructed to clear a particular pathogen may cause autoimmunity when cross-reacting with ignored self molecules. Alternatively, memory cells may induce an immune response of the wrong mode when cross-reacting with subsequent pathogens. To maximize the likelihood of responding to a wide variety of pathogens, it is also required that the immune system be sufficiently cross-reactive. By means of a probabilistic model, we show that these conflicting requirements are met optimally by a highly specific memory lymphocyte repertoire. This explains why the lymphocyte system that was built on a preserved functional innate immune system has such a high degree of specificity. Our analysis suggests that 1) memory lymphocytes should be more specific than naive lymphocytes and 2) species with small lymphocyte repertoires should be more vulnerable to both infection and autoimmune diseases.

Early recovery of CD4+ T lymphocytes in children on highly active antiretroviral therapy. Dutch study group for children with HIV infections.

INTRODUCTION: Regeneration of CD4+ T lymphocytes has been shown to be thymus-dependent in bone marrow transplant recipients and after intensive chemotherapy. The rate of CD4+ T cell regeneration is correlated positively with enlargement of the thymus, as shown on radiographs, and higher rates of CD4+ T lymphocyte regeneration were observed in children as compared with adults, consistent with thymic function diminishing with age. We hypothesized that in HIV infected patients CD4+ T cell recovery during highly active antiretroviral therapy (HAART) may also be thymus dependent. Therefore, repopulation of naive (CD45RA+), memory (CD45RO+) and total CD4+ T lymphocytes and total CD8+ T lymphocytes in peripheral blood was assessed in 13 HIV infected children during the initial 3 months of HAART. RESULTS: Significantly higher recovery rates of naive, memory and total CD4+ T cells were observed in children below the age of 3 years as compared with older children. Kinetics of total CD8+ T cells showed no relation to age. Moreover, recovery rates of naive CD4+ T cells in patients below 3 years of age were 10-40 fold higher as compared with previously reported naive CD4+ T cell recovery rates in adults on HAART. CONCLUSIONS: High recovery rates of naive, memory and total CD4+ T cells can be achieved in children below 3 years of age. Changes in CD8 counts did not correlate with age. These results indicate that regeneration of CD4+ T cells during HAART may be a thymus-dependent process.

T cell renewal rates, telomerase, and telomere length shortening.

Measurements on the average telomere lengths of normal human naive and memory T cells suggested that 1) naive and memory human T cells have similar division rates, and 2) that the difference between naive and memory cells reflects the degree of clonal expansion during normal immune reactions. Here we develop mathematic models describing how the population average of telomere length depends on the cell division rates of naive and memory T cells during clonal expansion and normal renewal. The results show that 1) telomeres shorten with twice the cell division rate, 2) that the conventional approach of estimating telomere length shortening per mean population doubling gives rise to estimates that are 39% larger than the "true" loss per cell division, 3) that naive and memory T cells are expected to shorten their telomeres at rates set by the division rate of the naive T cells only, i.e., irrespective of the division rate of memory T cells, 4) that the measured difference in the average telomere length between naive and memory T cells may largely reflect the difference in renewal rates between these subpopulations rather than the clonal expansion, and 5) that full telomerase compensation during clonal expansion is consistent with all data on the shortening of telomere length in, and between, naive and memory T cells. Thus we reconcile the apparent contradictions between the demonstrated difference in division rates between human naive and memory T cells and their similar rates of telomere shortening, and the demonstrated telomere shortening in the presence of telomerase activity.

Nonspatial visual attention explained by spatial attention plus limited storage.

The use of nonspatial attentional mechanisms in search tasks was investigated by presenting observers with stimuli that contained 4-12 elements located on a circle around the fixation point. The elements differed in one of six nonspatial 'dimensions', namely orientation, contrast, scale, number of cycles, 'shape', and place in the alphabet. The target element of the search task differed from trial to trial but was always presented to the observer as a nonspatial, visual cue. This cue was displayed either before the stimulus (precue) or after the stimulus (postcue). Whereas a precue creates optimal conditions for the use of nonspatial attentional mechanisms, a postcue precludes benefits from their use. The fact that performance was better in the case of precued stimuli than in the case of postcued stimuli indicates that observers employed nonspatial attentional mechanisms. In the final analysis, however, the effect of nonspatial attention reduces to spatial attention in combination with limited storage capacity.

Shape from shaded random surfaces.

The perception of surface relief from random shading patterns is measured by having observers adjust three-dimensional local probes, the projections of which are superimposed on the image. Three observers perform four settings of 91 probes on each of 14 images. These images are generated by calculating the Lambertian reflectance of a random superposition of elliptical Gaussian hills and valleys illuminated by a single distant light source as well as by ambient light. Neither the surface reflectance equation nor the light source direction is conveyed to our observers in any way. Mathematically, this "pure" shape-from-shading problem has highly non-unique solutions. Perception of a well-defined, stable shape therefore implies that the ambiguity is resolved, i.e. a gauge is fixed. We analyse the surface ambiguity or gauge freedom which is left unconstrained by pure shading information and we investigate possible ways of restricting it. Statistical analysis of the curl component of the field of probe settings reveals that the settings are significantly consistent with an underlying perceived surface. In spite of the large theoretical ambiguity in the stimuli, the settings are reproducible and show considerable inter-observer agreement. Even the correlation of the settings with the real surfaces is surprisingly large. If the settings are compared to the real surface normals, one finds a series of biases, the strongest of which is that the global surface slant is systematically underestimated, even in those cases where ending occluding contours or high-contrast luminance ridges, indicative of "almost" contours, are present in the image. Another bias then is that the corresponding rims on the surface are seen as roughly parallel to the picture plane.

Rapid segmentation of one-dimensional noise textures across borders.

We investigated the segmentation of texture pairs that were samples of one-dimensional binary visual noise. The stimulus consisted of an array of 5 x 8 squares separated by two-dimensional noise borders of varying width. The squares were filled with vertical black or white stripes of random width. The task was to detect the presence of a target square which differed from the squares above and below in one of three possible ways: the target pattern was either a contrast inverted copy or a horizontal translation of the pattern in the vertically adjacent squares, or else an independent realization of the noise. The binary noise in the textures was sequentially high-pass filtered to preclude the use of coarse-scale receptive fields and minimize the presence of sparse, extended "features". The target could be detected reliably within 100 msec even when the border width was larger than the maximal stripe width. The border width at threshold saturated for longer presentation times. Our results show that the microstructure of the patterns, i.e. information on the scale of the linewidth in the patterns, is not used directly, even though it contains most of the signal energy and is objectively the most reliable cue to the segmentation.

Motion transparency and coherence in plaids: the role of end-stopped cells.

Humans do not confound the motion of shadows cast upon a surface with the motion of the surface itself, although schemes that propose recombination of orientation-selective motion signals into a rigid motion percept of two-dimensional patterns would predict that they should do so. We propose a simple scheme that avoids recombination and instead attributes perception of two-dimensional pattern motion to the activation of orientation-selective end-stopped units that operate on the logarithm of the luminance. The proposed units respond to the change of contrast along a line, which typically occurs at an intersection. They are not active, however, when a shadow border intersects the edge of an object, because contrast does not change along either of these edges. Thus, end-stopped units signal the motion of transparent intersections weakly or not at all, and the independent motions of the shadow border and the object prevail. We tested two implications of this scheme, using plaids with variable intersection luminance. First, when the intersection luminance was such that it kept the contrast along the intersecting lines nearly constant, the sensitivity for the rigid plaid's direction of motion was minimal, and the sliding motion of the components prevailed. This occurred for light bars on dark backgrounds and for dark bars on light backgrounds. Thus, the effect of the intersection's luminance on the balance between the percepts of rigid-plaid motion and the motion of sliding components was independent of contrast inversion of bar and background. Secondly, when thin lines with the same luminance as the bars covered the borders of the intersection, the intersection's luminance did not affect the rigid-plaid motion percept very much, even when it corresponded to a transparent intersection. This indicates that, when the edges of the intersection and those of the bars were not collinear, the nulling of the end-stopped units did not occur. This result is in line with physiological studies, which showed that the response of an end-stopped cell to a line is only partially inhibited when a similar line is presented non-collinearly with the first in the inhibitory end-zone of its receptive field. Our results are consistent with a scheme in which a second stage of motion detectors combines signals of orientation-selective end-free and orientation-selective end-stopped units for perception of the rigid motion of two-dimensional patterns.

The role of early mechanisms in motion transparency and coherence.

Perceptual dissociation of moving plaid patterns into independently moving bar gratings occurs most readily when the grating signals are combined as if the bars were semi-transparent objects (Ramachandran, V. S. (1990) in: AI and the Eye. Wiley, Chichester, pp. 21-77. Stoner, G. R., Albright, T. D. and Ramachandran, V. S. (1990) Nature 344, 153-155). These and other examples of motion transparency are exploited to constrain the set of viable models for human motion processing. For example, one may exclude any fixed recombination of local motion signals into a plaid motion signal. Broad classes of linear and non-linear mechanisms for tracking blobs, corners, and other unambiguous plaid motion cues can also be ruled out because they fail to reproduce the experimental results even qualitatively. The shifting balance between the 'coherent plaid' and 'sliding gratings' percepts are attributed to processing stages before any integration or combination of local motion signals. The first essential stage is a roughly logarithmic nonlinearity before orientation filtering. In general, the resulting cross-products (at the intersections) code unambiguously for the true plaid motion vector, but these signals will be nulled for multiplicatively combined plaid components. Supporting evidence for this idea is obtained in our measurements of detection thresholds for the 'plaid' and 'sliding' percepts. The essential element of the second stage consists of 'end-stop' cells which detect the nullable intersection signal, and so produce a plaid-motion signal with the required characteristics. Finally, it is argued that the ecological role of the proposed mechanism lies in the ability to handle movement of patterned objects in lighting conditions dominated by complicated cast shadows.

Chaos or noise in EEG signals; dependence on state and brain site.

EEG signals have been considered to result either from random processes or to be generated by non-linear dynamic systems exhibiting chaotic behaviour. In the latter case, the system may behave as a deterministic chaotic attractor. The complexity of the attractor can be characterized by the correlation dimension that can be computed from one signal generated by the system. A new procedure was developed and applied in order to test whether the correlation dimension, calculated from an EEG epoch, may correspond to a chaotic attractor or to a random process. This procedure was applied to EEG signals recorded from different sites of the limbic cortex of the rat during different states: wakeful rest, locomotion and in the course of an epileptic seizure induced by kindling. The signals recorded during the first two states had high dimensions and could not be distinguished from random noise. However, during an epileptic seizure the correlation dimension became low (between 2 and 4) indicating that in this state the networks behave as chaotic systems. A low correlation dimension appeared at different times and brain sites during an epileptic seizure. These results show that the computation of the correlation dimension may be useful in order to obtain insight into the dynamics of the propagation of an epileptic seizure in the brain.