Predicting Amyloid-ß Levels in Amnestic Mild Cognitive Impairment Using Machine Learning Techniques.

BACKGROUND: Amyloid-ß positivity (Aß+) based on PET imaging is part of the enrollment criteria for many of the clinical trials of Alzheimer's disease (AD), particularly in trials for amyloid-targeted therapy. Predicting Aß positivity prior to PET imaging can decrease unnecessary patient burden and costs of running these trials. OBJECTIVE: The aim of this study was to evaluate the performance of a machine learning model in estimating the individual risk of Aß+ based on gold-standard of PET imaging. METHODS: We used data from an amnestic mild cognitive impairment (aMCI) subset of the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort to develop and validate the models. The predictors of Aß status included demographic and ApoE4 status in all models plus a combination of neuropsychological tests (NP), MRI volumetrics, and cerebrospinal fluid (CSF) biomarkers. RESULTS: The models that included NP and MRI measures separately showed an area under the receiver operating characteristics (AUC) of 0.74 and 0.72, respectively. However, using NP and MRI measures jointly in the model did not improve prediction. The models including CSF biomarkers significantly outperformed other models with AUCs between 0.89 to 0.92. CONCLUSIONS: Predictive models can be effectively used to identify persons with aMCI likely to be amyloid positive on a subsequent PET scan.

The Magnitude, But Not the Sign, of MT Single-Trial Spike-Time Correlations Predicts Motion Detection Performance.

Spike-time correlations capture the short timescale covariance between the activity of neurons on a single trial. These correlations can significantly vary in magnitude and sign from trial to trial, and have been proposed to contribute to information encoding in visual cortex. While monkeys performed a motion-pulse detection task, we examined the behavioral impact of both the magnitude and sign of single-trial spike-time correlations between two nonoverlapping pools of middle temporal (MT) neurons. We applied three single-trial measures of spike-time correlation between our multiunit MT spike trains (Pearson's, absolute value of Pearson's, and mutual information), and examined the degree to which they predicted a subject's performance on a trial-by-trial basis. We found that on each trial, positive and negative spike-time correlations were almost equally likely, and, once the correlational sign was accounted for, all three measures were similarly predictive of behavior. Importantly, just before the behaviorally relevant motion pulse occurred, single-trial spike-time correlations were as predictive of the performance of the animal as single-trial firing rates. While firing rates were positively associated with behavioral outcomes, the presence of either strong positive or negative correlations had a detrimental effect on behavior. These correlations occurred on short timescales, and the strongest positive and negative correlations modulated behavioral performance by ∼9%, compared with trials with no correlations. We suggest a model where spike-time correlations are associated with a common noise source for the two MT pools, which in turn decreases the signal-to-noise ratio of the integrated signals that drive motion detection.SIGNIFICANCE STATEMENT Previous work has shown that spike-time correlations occurring on short timescales can affect the encoding of visual inputs. Although spike-time correlations significantly vary in both magnitude and sign across trials, their impact on trial-by-trial behavior is not fully understood. Using neural recordings from area MT (middle temporal) in monkeys performing a motion-detection task using a brief stimulus, we found that both positive and negative spike-time correlations predicted behavioral responses as well as firing rate on a trial-by-trial basis. We propose that strong positive and negative spike-time correlations decreased behavioral performance by reducing the signal-to-noise ratio of integrated MT neural signals.

The cerebellum and visual perceptual learning: evidence from a motion extrapolation task.

Visual perceptual learning is widely assumed to reflect plastic changes occurring along the cerebro-cortical visual pathways, including at the earliest stages of processing, though increasing evidence indicates that higher-level brain areas are also involved. Here we addressed the possibility that the cerebellum plays an important role in visual perceptual learning. Within the realm of motor control, the cerebellum supports learning of new skills and recalibration of motor commands when movement execution is consistently perturbed (adaptation). Growing evidence indicates that the cerebellum is also involved in cognition and mediates forms of cognitive learning. Therefore, the obvious question arises whether the cerebellum might play a similar role in learning and adaptation within the perceptual domain. We explored a possible deficit in visual perceptual learning (and adaptation) in patients with cerebellar damage using variants of a novel motion extrapolation, psychophysical paradigm. Compared to their age- and gender-matched controls, patients with focal damage to the posterior (but not the anterior) cerebellum showed strongly diminished learning, in terms of both rate and amount of improvement over time. Consistent with a double-dissociation pattern, patients with focal damage to the anterior cerebellum instead showed more severe clinical motor deficits, indicative of a distinct role of the anterior cerebellum in the motor domain. The collected evidence demonstrates that a pure form of slow-incremental visual perceptual learning is crucially dependent on the intact cerebellum, bearing the notion that the human cerebellum acts as a learning device for motor, cognitive and perceptual functions. We interpret the deficit in terms of an inability to fine-tune predictive models of the incoming flow of visual perceptual input over time. Moreover, our results suggest a strong dissociation between the role of different portions of the cerebellum in motor versus non-motor functions, with only the posterior lobe being responsible for learning in the perceptual domain.

Selective tuning for contrast in macaque area V4.

Visually responsive neurons typically exhibit a monotonic-saturating increase of firing with luminance contrast of the stimulus and are able to adapt to the current spatiotemporal context by shifting their selectivity, therefore being perfectly suited for optimal contrast encoding and discrimination. Here we report the first evidence of the existence of neurons showing selective tuning for contrast in area V4d of the behaving macaque (Macaca mulatta), i.e., narrow bandpass filter neurons with peak activity encompassing the whole range of visible contrasts and pronounced attenuation at contrasts higher than the peak. Crucially, we found that contrast tuning emerges after a considerable delay from stimulus onset, likely reflecting the contribution of inhibitory mechanisms. Selective tuning for luminance contrast might support multiple functions, including contrast identification and the attentive selection of low contrast stimuli.

Topography of the motion aftereffect with and without eye movements.

Although a lot is known about various properties of the motion aftereffect (MAE), there is no systematic study of the topographic organization of MAE. In the current study, first we provided a topographic map of the MAE to investigate its spatial properties in detail. To provide a fine topographic map, we measured MAE with small test stimuli presented at different loci after adaptation to motion in a large region within the visual field. We found that strength of MAE is highest on the internal edge of the adapted area. Our results show a sharper aftereffect boundary for the shearing motion compared to compression and expansion boundaries. In the second experiment, using a similar paradigm, we investigated topographic deformation of the MAE area after a single saccadic eye movement. Surprisingly, we found that topographic map of MAE splits into two separate regions after the saccade: one corresponds to the retinal location of the adapted stimulus and the other matches the spatial location of the adapted region on the display screen. The effect was stronger at the retinotopic location. The third experiment is basically replication of the second experiment in a smaller zone that confirms the results of previous experiments in individual subjects. The eccentricity of spatiotopic area is different from retinotopic area in the second experiment; Experiment 3 controls the effect of eccentricity and confirms the major results of the second experiment.