239 research outputs found
Sort by
Bayesian local directional accuracy of a directionally selective Ganglion-cell ensemble
No full textOn-Off directionally selective ganglion cells (DSGC) of the rabbit retina send
information about the direction of motion to the rest of the brain. Each of these cells
responds best for motions in a preferred direction. There are four types of DSGC, each
type preferring motions along one Cartesian axis (up, down, left or right). Every point in
visual space is viewed by one cell of each type. We have measured the distribution of
responses of DSGC as a function of contrast and direction of motion. With this
information, and knowing the distribution of contrasts and directions of motion in natural
images, we can apply Bayesian Analysis to infer the possible local directional accuracy
of a DSGC ensemble's output. We applied Bayes Theorem:
P[c,f|r] =
P[r|c,f]P[c,f]
P[r]
where r, c, and f are response, contrast, and direction of motion respectively.
We performed both Maximum-Likelihood and Maximum-a-Posteriori analyses. The
former took only the biology into account (i.e., our measurements), while the latter
assumed an exponential distribution for contrasts (Balboa & Grzywacz, 2000; Tadmor &
Tolhurst, 2000) and a homogeneous distribution of directions of motion (Balboa &
Grzywacz, unpublished observations). We found that a local DSGC ensemble could
provide highly accurate directional estimates, with RMS errors of less than 3° for stimuli
with contrasts of 100%. Moreover, an ensemble could provide directional estimates with
less than10° RMS error for stimuli with contrasts of only 15% (the mean contrast in
natural images). Interestingly, including the contrast and direction-of-motion priors did
not improve performance significantly. This is because the Maximum-Likelihood
estimate only fails when the biological system is uncertain. DSGCs have relatively low
noise and therefore, low uncertainty. It remains to be seen whether other prior
information, such as the distribution of speeds or spatial frequencies, can improve the
DSGC system's directional accuracy
Neural control of behavior in a neuromechanical salamander simulation
No full textWith arguably the simplest vertebrate nervous system, the salamander is a model organism used to study basic issues in vertebrate neuroscience. We are using a neuromechanical simulation program to investigate behavioral consequences of the amphibian's neural organization. The system's Central Pattern Generators (CPGs)produce biologically plausible spinal waveforms, thus producing characteristic walking and swimming movement. The system has been used to investigate prey approach behavior with the realistic mechanical simulation and a simplified visual system model.
We are using this model to investigate the relative contributions of sensory feedback and direct CPG control on head stabilization during locomotion. We will use computer vision techniques to determine, from video obtained under controlled conditions, precise kinematic parameters of the locomotion gaits, particularly head movement parameters.
These kinematic parameters will constrain the simulation. From a current study, it appears likely that the pattern generators play the predominant role in head stabilization.
We will describe an approach to meeting the kinematic constraints using the CPG model. A somewhat related effort studies a model of the amphibian medial pallium (MP) -- a
likely homology to the hippocampal formation. Recent work on the salamander has produced neuroanatomical results that will constrain an older model of the toad MP that
explained intricate patterns of habituation and ishabituation to prey-like stimuli.
Investigators have hypothesized that hippocampus homologies are involved in evolutionarily conserved patterns of vertebrate behavior – e.g., spatial memory involved in navigation and perhaps other forms of territorial behavior. This computational investigation analyzes such a hypothesis in the context of a model which has demonstrated biological plausibility at other levels of analysis. The fusion of these disparate investigations will be a unified, modular, biologically plausible model of an autonomous agent successfully interacting with a complex environment in order to satisfy biological drives – a computational neuroethological model. Questions about the interfaces of these systems will be discussed
Computational subunits in thin dendrites of pyramidal cells
No full textThe thin basal and oblique dendrites of cortical pyramidal neurons receive most of the cells' synaptic input, but their integrative properties remain uncertain. Previous studies have most often reported global linear or sublinear summation. An alternative view,supported by biophysical modeling studies, holds that thin dendrites provide a layer of independent computational 'subunits' that sigmoidally modulate their inputs prior to global summation. To distinguish these possibilities, we combined confocal imaging and dual-site focal synaptic stimulation of identified thin dendrites in rat neocortical
pyramidal neurons. We found that nearby inputs on the same branch summed sigmoidally, whereas widely separated inputs or inputs to different branches summed
linearly. This strong spatial compartmentalization effect is incompatible with a global
summation rule and provides the first experimental support for a two-layer "neural network" [The quotes are left in to refer to a standard architecture in the artificial neural
network field] model of pyramidal neuron thin-branch integration. Our findings could have important implications for the computing and memory-related functions of cortical
tissue
Estimation of linear and nonlinear spatial receptive fields from natural Images
No full textPlease see attached pd
Use of correlation to assess the signal content of patterns of spike activity reflecting the population response of relays of the central auditorimotor system of cats
No full textPlease see attached PD
Disentangling topdown from bottom up influences on attentional allocation in dynamic scenes
No full textMotivation:
Attentional allocation is determined by the interplay between bottom-up and top-down influences. Here we try to quantify the relative contributions of different influences on attentional allocation in dynamic scenes, as well as examine how they change over time.
Methods:
In order to manipulate the availability of top-down influences on attentional allocation, heterogeneous video clips were cut into clippets (M=2s), which were scrambled and re-assembled into MTV-style clips. Two groups of 8 Subjects each were instructed to "follow the main actors and
actions". One group viewd the original stimuli while the other group viewd the MTV-style clips. Eye positions were recorded using an ISCAN eye-tracker (240Hz, yielding a total of more than a million samples for each group), and
segmented into saccades, blinks, and fixation/smooth pursuit periods. A saliency-based model of attention capture (Itti & Koch 2000) was used to probe the relative contribution of bottom-up influences on attentional
allocation based on a novel performance metric - Chance-Adjusted Saliency Accumometric (CASA). CASA values were computed based on the weighted sum of differences between normalized saliency at human vs. random saccade targets.
Results:
Total CASA based on the full saliency model was 6% higher in the MTV group compared to the original group. In both original and MTV groups, CASA based on either motion or flicker features alone was ~95% of the CASA
based on the full saliency model. CASA based on either color, intensity, or orientation features alone was ~66% of the full model CASA. Generally, CASA values for earlier saccades after stimulus onset (clip or clippet start)
were higher than for later saccades, but tapered off and flactuated around a fairly high value after the first several saccades.
Conclusions:
The 6% CASA difference between the original and MTV groups
shows that eliminating visual context beyond the first ~2s of viewing barely increased the overall relative weight of bottom-up influences on attentional allocation. Our results imply that the relative weight of top-down influences on attentional allocation in dynamic scenes does not increase with viewing time (beyond the first ~2s). We also found that
either motion or flicker are ~150% stronger than either color, intensity, or orientation as bottom-up attractors of attention
Parameters for Modeling Behavioral Effects of Focal Stimulation
No full textA major research method in neuroscience is manipulation of relative regional activation by focal stimulation, followed by the observation of behavioral effects. Electrical methods include direct current stimulation (Antal et al., 2004), and transcranial magnetic stimulation (Fierro et al., 2000). Perceptually-based methods include tachistoscopic (Zarate et al., 2000), partially-obscured contact lens (Levick et al., 1993), haptic (Bassel & Schiff, 2001), auditory (Drake & Sobrero, 1987), or motor (McCourt et al., 2001; Schiff et al., 1998) orientation of
attention. Biofeedback is also effective for the voluntary control of regionalcerebral activation (Pulvermuller, Mohr, Schleichert, & Veit, 2000; Rosenfeld,Cha, Blair, & Gotlib,1995; Schwartz, Davidson, & Pugash, 1976).
A computational model linking these manipulations with the resulting cognitive and affective behavior needs to account for at least the following three parameters: a) an individual difference in variability of functions among regions (Chiarello, Kacinik, Manowitz, Otto, & Leonard, 2004; Hellige, Bloch, Cowin, Eng, Eviatar, & Sergent,1994; Neubauer, Grabner, Freudenthaler, Beckmann, &
Guthke, 2004), b) an individual difference in reactivity to the manipulation (Wexler, Schwartz, Warrenburg, Servis, & Tarlatzis, 1986; Wheeler, Davidson, & Tomarken, 1993), and c) the effective strength of the manipulation.
These parameters will probably be multiplicative, but perhaps with logarithmic or power factors. A tentative computational formula is:
dQ = a[1 + b(c)] where dQ is delta or differential change in behavior as a result of the manipulation. An example of such a change is increase or decrease in risk taking because of manipulated stimulation of a region (Drake, 1985). This phenomenon is replicated in lesion studies (Miller & Milner, 1985), but the datafitting
power of the formula and the relative contribution of each variable await parametric empirical testing.
Support for this submission was provided in part by a grant from the National Institute on Drug Abuse to the USC Institute for Prevention Research (P50-DA16094)
Neural Correlate of Object-Based Selection in Area V4
No full textSingle unit studies of attention in monkeys have identified competitive circuits in extrastriate cortex that could mediate selection of one stimulus over another. While these
studies show that attention operates by resolving competition, they used stimuli atseparate locations, confounding selection of objects with selection of spatial locations. To resolve this, we recorded responses of V4 neurons to two spatially superimposed transparent surfaces, one of which was delayed in onset. The surfaces were defined by patterns of dots that rotated rigidly around a common center. One set of dots was of the neuron's preferred color and the other was of an isoluminant non-preferred color. Human psychophysics using the same type of stimuli found that the delayed onset of one surface exogenously cues attention to that surface and suppresses processing of the other surface for several hundred milliseconds. Consistent with this, neurons in area V4 were
preferentially driven by the delayed surface.
Using superimposed surfaces ruled out spatial selection. But is this selection object-based? If it is, the selection should survive moving the superimposed surfaces
through space. When the appearance of one of the two surfaces was delayed outside the neuron's receptive field and both surfaces then moved into the RF, the pair response was still preferentially driven by the delayed surface.
Neurophysiological and functional imaging studies have shown that endogenously directing attention to the color or motion of a stimulus preferentially processes it
throughout the visual field. We tested for feature-based selection by using placing two surfaces within the RF and two outside of the RF. When the delayed surface appeared
within the RF, the results were similar to the first experiment, i.e. the delayed surface was preferentially processed. If this effect were the result of global color-based selection, thenthe same effect should be seen when the delayed surface appeared outside the RF. This
effect was not seen, hence the selection was not of the color of the surface but of the surface itself.
These results show that competitive circuits in V4 are not limited to mediating competition between spatial locations, but also select objects. These circuits are a likely
neural substrate for object-based attention
What crowds a letter in the periphery?
No full textIn the periphery, nearby letters (flankers) can impair the
identification of a target letter, but white-noise patches or gratings do not (Palomares et al., ARVO99), suggesting that letter crowding is unlike contrast masking (Chung, Levi, Legge, 2001). The aim of this study was to systematically test different types of flankers that may induce letter crowding in the periphery. Subjects identified a target letter presented at 5 deg inferior to fixation. Letters were in Times New Roman, of size 0.4 log
units above the subject's acuity. Six conditions were ested: target letter alone, flanked by same-polarity letter flankers, opposite-polarity letter flankers, "letter"- noise patches, or white-noise patches at 1 x-height spacing,
or flanked by same-polarity letter flankers at 2 x-height spacing. The letternoise patch was created by scrambling the phase spectrum of a letter but retaining its power spectral density (PSD). Moreover, both types of noise
had the same contrast energy and bounding box as the letter they replaced. Flankers were at 20% contrast. Threshold contrast of the target letter was determined at 79% correct. Average threshold contrast (across 3 Ss) for the
target-alone condition was 12.8% (+/-0.04 log units). No significant threshold elevation was observed for white-noise patches or letter flankers at 2x spacing. All other conditions led to significant crowding (avg. threshold
elevation in log units - letter: 0.35, letter noise: 0.26, reverse contrast: 0.21). Threshold elevation between letter and letter noise flankers was not significant for 2 out of the 3 Ss. Letter-noise patches caused significant
crowding, while white-noise patches did not. Spatial frequency distribution thus seems to play a major role in letter crowding. Phase, however, which defines the visual form of the flankers, had only a limited role. We speculate
that a main component of letter crowding may be noise masking, with noise being induced by and having a similar PSD to the flankers