PubMed Health⌕ Search

Biomedical subjects

B J Richmond

Publications and source records attributed to B J Richmond.

At least 19 recordsLinked to original sources
PubMed1997

Neuronal codes: reading them and learning how their structure influences network organization.

Our investigations of the primate visual system show that neuronal responses carry information in a multi-dimensional code that is superimposed onto the response envelope in a slow time varying fashion. The precision of timing is 30 ms or more. In primary visual cortex response latency and response strength are largely independent, with latency more closely coding contrast or visibility and strength more closely coding stimulus orientation, or perhaps shape. Adjacent neurons in both V1 and inferior temporal cortex share only about 10% of their stimulus-related information, which we demonstrate to be consistent with the idea that cortical layers were organized to minimize information loss.

Animals↗
PubMed1996

Neural signals in the monkey ventral striatum related to motivation for juice and cocaine rewards.

1. The results of neuropsychological, neuropharmacological, and neurophysiological experiments have implicated the ventral striatum in reward-related processes. We designed a task to allow us to separate the effects of sensory, motor, and internal signals so that we could study the correlation between the activity of neurons in the ventral striatum and different motivational states. In this task, a visual stimulus was used to cue the monkeys as to their progress toward earning a reward. The monkeys performed more quickly and with fewer mistakes in the rewarded trials. After analyzing the behavioral results from three monkeys, we recorded from 143 neurons from two of the monkeys while they performed the task with either juice or cocaine reward. 2. In this task the monkey was required to release its grip on a bar when a small visual response cue changed colors from red (the wait signal) to green (the go signal). The duration of the wait signal was varied randomly. The cue became blue whenever the monkey successfully responded to the go signal within 1 s of its appearance. A reward was delivered after the monkey successfully completed one, two, or three trials. The schedules were randomly interleaved. A second visual stimulus that progressively brightened or dimmed signaled to the monkeys their progress toward earning a reward. This discriminative cue allowed the monkeys to judge the proportion of work remaining in the current ratio schedule of reinforcement. Data were collected from three monkeys while they performed this task. 3. The average reaction times became faster and error rates declined as the monkeys progressed toward completing the current schedule of reinforcement and thereby earning a reward, whereas the modal reaction time did not change. As the duration of the wait period before the go signal increased, the monkeys reacted more quickly but their error rates scarcely changed. From these results we infer that the effects of motivation and motor readiness in this task are generated by separate mechanisms rather than by a single mechanism subserving generalized arousal. 4. The activity of 138 ventral striatal neurons was sampled in two monkeys while they performed the task to earn juice reward. We saw tonic changes in activity throughout the trials, and we saw phasic activity following the reward. The activity of these neurons was markedly different during juice-rewarded trials than during correctly performed trials when no reward was forthcoming (or expected). The responses also were weakly, but significantly, related to the proximity of the reward in the schedules requiring more than one trial. 5. The monkeys worked to obtain intravenous cocaine while we recorded 62 neurons. For 57 of the neurons, we recorded activity while the monkeys worked in blocks of trials during which they self-administered cocaine after blocks during which they worked for juice. Although fewer neurons responded to cocaine than to juice reward (19 vs. 33%), this difference was not significant. The neuronal response properties to cocaine and juice rewards were independent; that is, the responses when one was the reward one failed to predict the response when the other was the reward. In addition, the neuronal activity lost most of its selectivity for rewarded trials, i.e, the activity did not distinguish nearly as well between cocaine and sham rewards as between juice and sham rewards. 6. Our results show that mechanisms by which cocaine acts do not appear to be the same as the ones activated when the monkeys were presented with an oral juice reward. This finding raises the intriguing possibility that the effects of cocaine could be reduced selectively without blocking the effects of many natural rewards.

Animals↗
PubMed1996

Latency: another potential code for feature binding in striate cortex.

1. We recorded the responses of 37 striate cortical complex cells in fixating monkeys while presenting a set of oriented stimuli that varied in contrast. 2. The two response parameters of strength and latency can be interpreted as a code: the strength defines the stimulus form (here the orientation), and the latency is more a function of the stimulus contrast. 3. Synchronization based on latency could make a strong contribution to the process of organizing the neural responses to different objects, i.e., binding.

Animals↗
PubMed1995

Information flow and temporal coding in primate pattern vision.

We perform time-resolved calculations of the information transmitted about visual patterns by neurons in primary visual and inferior temporal cortices. All measurable information is carried in an effective time-varying firing rate, obtained by averaging the neuronal response with a resolution no finer than about 25 ms in primary visual cortex and around twice that in inferior temporal cortex. We found no better way for a neuron receiving these messages to decode them than simply to count spikes for this long. Most of the information tends to be concentrated in one or, more often, two brief packets, one at the very beginning of the response and the other typically 100 ms later. The first packet is the most informative part of the message, but the second one generally contains new information. A small but significant part of the total information in the message accumulates gradually over the entire course of the response. These findings impose strong constraints on the codes used by these neurons.

Animals↗
PubMed1994

Decoding cortical neuronal signals: network models, information estimation and spatial tuning.

We have studied the encoding of spatial pattern information by complex cells in the primary visual cortex of awake monkeys. Three models for the conditional probabilities of different stimuli, given the neuronal response, were fit and compared using cross-validation. For our data, a feed-forward neural network proved to be the best of these models. The information carried by a cell about a stimulus set can be calculated from the estimated conditional probabilities. We performed a spatial spectroscopy of the encoding, examining how the transmitted information varies with both the average coarseness of the stimulus set and the coarseness differences within it. We find that each neuron encodes information about many features at multiple scales. Our data do not appear to allow a characterization of these variations in terms of the detection of simple single features such as oriented bars.

Animals↗
PubMed1993

How independent are the messages carried by adjacent inferior temporal cortical neurons?

There are at least three possibilities for encoding information in a small area of cortex. First, neurons could have identical characteristics, thus conveying redundant information; second, neurons could give different responses to the same stimuli, thus conveying independent information; or third, neurons could cooperate with each other to encode more information jointly than they do separately, that is, synergistically. We recorded from 28 pairs of neurons in inferior temporal cortex of behaving rhesus monkeys. Each pair was recorded from a single microelectrode. Both the magnitude and the temporal modulation of the responses were quantified. We separated the responses into signal (average response to each stimulus) and noise (deviation of each response from the average). Linear regression showed that an average of only 18.7% of the magnitude of the signal carried by one neuron could be predicted from the magnitude of the other, and only 22.0% could be predicted by including the temporal modulation. For the noise, the figures were 5.5% and 6.3%, respectively, even less than for the signal. Information theoretic analysis shows that the pairs of neurons we studied carried an average of 20% redundant information. However, even this relatively small amount of redundancy places a severe upper limit on the information that can be transmitted by a neuronal pool. A pool of neurons for which each pair is mutually redundant to extent y can only carry a maximum of 1/y, here five times, as much information as one neuron alone. Information theoretic analysis gave no evidence for the presence of information as a function of both neurons considered together, that is, synergistic codes. Cross-correlation showed that at least 61% of the neuronal pairs shared connections in some manner. Given these shared connections, if adjacent neurons had had identical characteristics, then the noise on the outputs of these neurons would have been highly correlated, and it would not be possible to separate the signal and noise. The severe impact of correlated noise and information redundancy leads us to propose that the processing carried out by these neurons evolved both to provide a rich description of many stimulus properties and simultaneously to minimize the redundancy in a local group of neurons. These two principles appear to be a major constraint on the organization of inferior temporal, and possibly all, cortex.

Animals↗
PubMed1992

Medial malleolar stress fractures in seven patients: review of the clinical and imaging features.

Stress fractures of the medial malleolus were discovered in seven patients, five male and two female subjects aged 16-34 years. All except one were involved in running and jumping athletic activities. Gradual onset of pain over the medial malleolus occurred with repetitive activity. Focal intense increased uptake in the medial malleolus was present on bone scans. Conventional radiography and computed tomography demonstrated the presence of subtle fissures at the junction of the medial malleolus and the tibial plafond, and well-circumscribed lytic lesions were also seen in three patients. Two patients developed a complete fracture of the medial malleolus. Stress fractures of the medial malleolus should be suspected in patients involved in athletic and/or unusual activities who have experienced persistent and unexplained pain over the medial malleolus. Bone scans and radiographs should be obtained for diagnostic purposes in these patients.

Adolescent↗
PubMed1992

Role of inferior temporal neurons in visual memory. I. Temporal encoding of information about visual images, recalled images, and behavioral context.

1. Lesions of the inferior temporal (IT) cortex selectively hamper monkeys in tasks requiring visual memory. A system that recognizes images must be able to encode a current stimulus, recall the code of a previous stimulus, compare the codes of the two stimuli, and make a decision on the basis of the outcome of the comparison. Therefore, IT neurons must be involved in at least one of these processes. To determine the specific role of IT neurons in visual memory, we measured the information conveyed in the neuronal responses about current patterns, recalled patterns, and behavioral context. 2. Two monkeys were trained to perform a sequential matching task using a set of 32 black and white Walsh patterns. In the course of an experiment, each pattern was presented repeatedly in sample, match, and nonmatch behavioral contexts. While the monkeys were performing the task, we recorded the activity of 76 neurons from area TE of IT. The neuronal responses to the stimuli were converted to spike density functions, and the resultant waveforms were quantified using their principal components. The relationships between the responses and the stimuli were studied using analysis of variance and information theory. 3. The analysis of variance was applied to the neuronal response waveforms using the context (sample, match, or nonmatch) and the patterns of the stimuli as independent variables and the spike count or the coefficients of the principal components as the dependent variables. We found that the waveforms of most neurons were significantly modulated by both the pattern and the context of the stimulus presentation. 4. We also analyzed the stimulus-response relationships using information theory. The input codes were based on the pattern and context of the stimuli, and the output codes were based on the spike count or the principal components of the responses. The neuronal response waveforms were found to convey significant amounts of information about both the pattern and context of the stimuli. Transmitted information was greatest when the response of a neuron was interpreted as a message about the combination of pattern and context. Nevertheless, there was information about context independent of pattern and vice versa. 5. We also used information theory to determine whether the neuronal responses to the second, or test, stimulus conveyed information about the pattern of the first, or sample, stimulus. The input codes were based on the patterns of the sample stimuli, and the output codes were based on the responses to the nonmatch test stimuli.(ABSTRACT TRUNCATED AT 400 WORDS)

Analysis of Variance↗
PubMed1992

Role of inferior temporal neurons in visual memory. II. Multiplying temporal waveforms related to vision and memory.

1. In the companion paper we reported on the activity of neurons in the inferior temporal (IT) cortex during a sequential pattern matching task. In this task a sample stimulus was followed by a test stimulus that was either a match or a nonmatch. Many of the neurons encoded information about the patterns of both current and previous stimuli in the temporal modulation of their responses. 2. A simple information processing model of visual memory can be formed with just four steps: 1) encode the current stimulus; 2) recall the code of a remembered stimulus; 3) compare the two codes; 4) and decide whether they are similar or different. The analysis presented in the first paper suggested that some IT neurons were performing the comparison step of visual memory. 3. We propose that IT neurons participate in the comparison of temporal waveforms related to vision and memory by multiplying them together. This product could form the basis of a crosscorrelation-based comparison. 4. We tested our hypothesis by fitting a simple multiplicative model to data from IT neurons. The model generated waveforms in separate memory and visual channels. The waveforms arising from the two channels were then multiplied on a point by point basis to yield the output waveform. The model was fitted to the actual neuronal data by a gradient descent method to find the best fit waveforms that also had the lowest total energy. 5. The multiplicative model fit the neuronal responses quite well. The multiplicative model made consistently better predictions of the actual response waveforms than did an additive model. Furthermore, the fit was better when the actual relationship between the responses and the sample and test stimuli were preserved than when that relationship was randomized. 6. We infer from the superior fit of the multiplicative model that IT neurons are multiplying temporally modulated waveforms arising from separate visual and memory systems in the comparison step of visual memory.

Animals↗
PubMed1992

Diagnostic efficacy of digitized images vs plain films: a study of the joints of the fingers.

Four hundred fifteen finger joints from 30 patients were evaluated for the presence of joint-space erosion, narrowing, and degenerative spurring on plain films, low-resolution digitized images (1024 x 840 bytes x 12 bit matrix), and high-resolution digitized images (2048 x 1680 bytes x 12 bit matrix). Three hundred four joints were abnormal. Low- and high-resolution digital images were displayed on a 1K x 1K monitor with the ability to change level, window, orientation, and brightness. Five radiologists interpreted images. The presence or absence of each abnormality was determined by consensus of two skeletal radiologists who did not otherwise participate in the study. Receiver-operating-characteristic analysis was used to obtain an area and a true-positive rate at a 0.10 false-positive rate for each interpreter. Randomized block analysis of variance with interpreters as blocks was used to compare areas and true-positive rates among imaging techniques for each type of abnormality; no statistically significant differences were found. In conclusion, the efficacy of display of digitized images on high- and low-resolution modes is not significantly different from that of plain films in the detection of erosions, joint-space narrowing, or degenerative spurring in small joints of the hands.

Analysis of Variance↗
PubMed1991

Concurrent processing and complexity of temporally encoded neuronal messages in visual perception.

The intrinsic neuronal code that carries visual information and the perceptual mechanism for decoding that information are not known. However, multivariate statistics and information theory show that neurons in four visual areas simultaneously carry multiple, stimulus-related messages by utilizing multiplexed temporal codes. The complexity of these temporal messages increases progressively across the visual system, yet the temporal codes overlap in time. Thus, visual perception may depend on the concurrent processing of multiplexed temporal messages from all visual areas.

Animals↗
PubMed1991

Unbiased measures of transmitted information and channel capacity from multivariate neuronal data.

Two measures from information theory, transmitted information and channel capacity, can quantify the ability of neurons to convey stimulus-dependent information. These measures are calculated using probability functions estimated from stimulus-response data. However, these estimates are biased by response quantization, noise, and small sample sizes. Improved estimators are developed in this paper that depend on both an estimate of the sample-size bias and the noise in the data.

Animals↗
PubMed1991

Task difficulty: ignoring, attending to, and discriminating a visual stimulus yield progressively more activity in inferior temporal neurons.

To study the influence of task difficulty on the stimulus-elicited responses of inferior temporal (IT) neurons, the stimulus-elicited responses of 64 neurons were recorded from IT cortex of three rhesus monkeys while they performed three behavioral tasks-an irrelevant-stimulus task, a stimulus detection task, and a stimulus discrimination task. The monkey could ignore the stimulus entirely in the irrelevant-stimulus task, was required only to detect stimulus dimming in the stimulus detection task, and was required to attend to specific properties of the stimulus in the discrimination task. The excitatory responses in the discrimination and stimulus detection tasks were larger than those in the irrelevant-stimulus task (61% and 33%, respectively, of the individual differences were significant), and excitatory responses in the discrimination task were larger than those in the detection task (49% of the individual differences reached significance). Twenty percent of the stimulus presentations elicited inhibitory responses that were followed by off-responses. The off-responses were modulated by the tasks in the same order as the excitatory on-responses. Assuming that the off-response strengths indicate the depth of the stimulus-induced inhibition, these results suggest that inhibitory responses were influenced across these tasks in a manner similar to the excitatory responses. When the neuronal responses were related to the difficulties of these tasks, both the response strength and errors were seen to be least during the irrelevant-stimulus task and greatest during the discrimination task. This relationship suggests that the visual responsiveness of IT neurons is related to the degree of attention the animal pays to the stimulus.(ABSTRACT TRUNCATED AT 250 WORDS)

Animals↗
PubMed1991

Application of an artificial neural network in radiographic diagnosis.

The description of 44 cases of bone tumors was used by an artificial neural network to rank the likelihood of 55 possible pathologic diagnoses. The performance of the artificial neural network was compared with the performance of experienced (3 or more years of radiology training) residents and inexperienced (less than 1 year of radiology training) residents. The artificial neural network was trained using descriptions of 110 radiographs of bone tumors with known diagnoses. The descriptions of a separate set of 44 cases were used to test the neural network. The neural network ranked 55 possible pathologic diagnoses on a scale from 1 to 55. Experienced and inexperienced residents also ranked the possible diagnoses in the same 44 cases. Inexperienced residents had a significantly lower mean proportion of diagnoses ranked first or second than did the neural network. Experienced residents had a significantly higher proportion of correct diagnoses ranked first than did the network. Otherwise, a significant difference between the performance of the network and experienced or inexperienced residents was not identified. These results demonstrate that artificial neural networks can be trained to classify bone tumors. Whether neural network performance in classification of bone tumors can be made accurate enough to assist radiologists in clinical practice remains an open question. These preliminary results indicate that further investigation of this technology for interpretation assistance is warranted.

Bone Neoplasms↗
PubMed1991

Interactive effects among several stimulus parameters on the responses of striate cortical complex cells.

1. Although neurons within the visual system are often described in terms of their responses to particular patterns such as bars and edges, they are actually sensitive to many different stimulus features, such as the luminances making up the patterns and the duration of presentation. Many different combinations of stimulus parameters can result in the same neuronal response, raising the problem of how the nervous system can extract information about visual stimuli from such inherently ambiguous responses. It has been shown that complex cells transmit significant amounts of information in the temporal modulation of their responses, raising the possibility that different stimulus parameters are encoded in different aspects of the response. To find out how much information is actually available about individual stimulus parameters, we examined the interactions among three stimulus parameters in the temporally modulated responses of striate cortical complex cells. 2. Sixteen black and white patterns were presented to two awake monkeys at each of four luminance-combinations and five durations, giving a total of 320 unique stimuli. Complex cells were recorded in layers 2 and 3 of striate cortex, with the stimuli centered on the receptive fields as determined by mapping with black and white bars. 3. An analysis of variance (ANOVA) was applied to these data with the three stimulus parameters of pattern, the luminance-combinations, and duration as the independent variables. The ANOVA was repeated with the magnitude and three different aspects of the temporal modulation of the response as the dependent variables. For the 19 neurons studied, many of the interactions between the different stimulus parameters were statistically significant. For some response measures the interactions accounted for more than one-half of the total response variance. 4. We also analyzed the stimulus-response relationships with the use of information theoretical techniques. We defined input codes on the basis of each stimulus parameter alone, as well as their combinations, and output codes on the basis of response strength, and on three measures of temporal modulation, also taken individually and together. Transmitted information was greatest when the response of a neuron was interpreted as a temporally modulated message about combinations of all three stimulus parameters. The interaction terms of the ANOVA suggest that the response of a complex cell can only be interpreted as a message about combinations of all three stimulus parameters.(ABSTRACT TRUNCATED AT 400 WORDS)

Analysis of Variance↗
PubMed1991

Lateral geniculate neurons in behaving primates. I. Responses to two-dimensional stimuli.

1. Using behaving monkeys, we studied the visual responses of single neurons in the parvocellular layers of the lateral geniculate nucleus (LGN) to a set of two-dimensional black and white patterns. We found that monkeys could be trained to make sufficiently reliable and stable fixations to enable us to plot and characterize the receptive fields of individual neurons. A qualitative examination of rasters and a statistical analysis of the data revealed that the responses of neurons were related to the stimuli. 2. The data from 5 of the 13 "X-like" neurons in our sample indicated the presence of antagonistic center and surround mechanisms and linear summation of luminance within center and surround mechanisms. We attribute the lack of evidence for surround antagonism in the eight neurons that failed to exhibit center-surround antagonism either to a mismatch between the size of the pixels in the stimuli and the size of the receptive field or to the lack of a surround mechanism (i.e., the type II neurons of Wiesel and Hubel). 3. The data from five other neurons confirm and extend previous reports indicating that the surround regions of X-like neurons can have nonlinearities. The responses of these neurons were not modulated when a contrast-reversing, bipartite stimulus was centered on the receptive field, which suggests a linear summation within the center and surround mechanisms. However, it was frequently the case for these neurons that stimuli of identical pattern but opposite contrast elicited responses of similar polarity, which indicates nonlinear behavior. 4. We found a wide variety of temporal patterns in the responses of individual LGN neurons, which included differences in the magnitude, width, and number of peaks of the initial on-transient and in the magnitude of the later sustained component. These different temporal patterns were repeatable and clearly different for different visual patterns. These results suggest that visual information may be carried in the shape as well as in the amplitude of the response waveform.

Animals↗