Catalogue Search | MBRL
Are you sure you want to remove the book from the shelf?
{{itemTitle}}
618
result(s) for
"Peters, Benjamin"
Sort by:
Context information supports serial dependence of multiple visual objects across memory episodes
Serial dependence is thought to promote perceptual stability by compensating for small changes of an object’s appearance across memory episodes. So far, it has been studied in situations that comprised only a single object. The question of how we selectively create temporal stability of several objects remains unsolved. In a memory task, objects can be differentiated by their to-be-memorized feature (content) as well as accompanying discriminative features (context). We test whether congruent context features, in addition to content similarity, support serial dependence. In four experiments, we observe a stronger serial dependence between objects that share the same context features across trials. Apparently, the binding of content and context features is not erased but rather carried over to the subsequent memory episode. As this reflects temporal dependencies in natural settings, our findings reveal a mechanism that integrates corresponding content and context features to support stable representations of individualized objects over time.
Visual cognition compensates for small changes in an object’s appearance to ensure its perceived continuity. We show that in situations with multiple objects, context features like color, temporal or spatial position are used as anchors to selectively integrate corresponding objects over time.
Journal Article
Can neural networks benefit from objectives that encourage iterative convergent computations? A case study of ResNets and object classification
Recent work has suggested that feedforward residual neural networks (ResNets) approximate iterative recurrent computations. Iterative computations are useful in many domains, so they might provide good solutions for neural networks to learn. However, principled methods for measuring and manipulating iterative convergence in neural networks remain lacking. Here we address this gap by 1) quantifying the degree to which ResNets learn iterative solutions and 2) introducing a regularization approach that encourages the learning of iterative solutions. Iterative methods are characterized by two properties: iteration and convergence. To quantify these properties, we define three indices of iterative convergence. Consistent with previous work, we show that, even though ResNets can express iterative solutions, they do not learn them when trained conventionally on computer-vision tasks. We then introduce regularizations to encourage iterative convergent computation and test whether this provides a useful inductive bias. To make the networks more iterative, we manipulate the degree of weight sharing across layers using soft gradient coupling. This new method provides a form of recurrence regularization and can interpolate smoothly between an ordinary ResNet and a “recurrent” ResNet (i.e., one that uses identical weights across layers and thus could be physically implemented with a recurrent network computing the successive stages iteratively across time). To make the networks more convergent we impose a Lipschitz constraint on the residual functions using spectral normalization. The three indices of iterative convergence reveal that the gradient coupling and the Lipschitz constraint succeed at making the networks iterative and convergent, respectively. To showcase the practicality of our approach, we study how iterative convergence impacts generalization on standard visual recognition tasks (MNIST, CIFAR-10, CIFAR-100) or challenging recognition tasks with partial occlusions (Digitclutter). We find that iterative convergent computation, in these tasks, does not provide a useful inductive bias for ResNets. Importantly, our approach may be useful for investigating other network architectures and tasks as well and we hope that our study provides a useful starting point for investigating the broader question of whether iterative convergence can help neural networks in their generalization.
Journal Article
Human-centred physical neuromorphics with visual brain-computer interfaces
Steady-state visual evoked potentials (SSVEPs) are widely used for brain-computer interfaces (BCIs) as they provide a stable and efficient means to connect the computer to the brain with a simple flickering light. Previous studies focused on low-density frequency division multiplexing techniques, i.e. typically employing one or two light-modulation frequencies during a single flickering light stimulation. Here we show that it is possible to encode information in SSVEPs excited by high-density frequency division multiplexing, involving hundreds of frequencies. We then demonstrate the ability to transmit entire images from the computer to the brain/EEG read-out in relatively short times. High-density frequency multiplexing also allows to implement a photonic neural network utilizing SSVEPs, that is applied to simple classification tasks and exhibits promising scalability properties by connecting multiple brains in series. Our findings open up new possibilities for the field of neural interfaces, holding potential for various applications, including assistive technologies and cognitive enhancements, to further improve human-machine interactions.
Steady state visually evoked potentials enable EEG vision-based brain computer interfaces. We show that it is possible to encode information distributed across 100+ frequencies. We encode images and perform simple physical computing classification tasks. Connecting more than one in brain in series improves the classification capability.
Journal Article
Perception of effort during an isometric contraction is influenced by prior muscle lengthening or shortening
PurposeFollowing a shortening or lengthening muscle contraction, torque produced in the isometric steady state is lower (residual torque depression; rTD) or higher (residual torque enhancement; rTE), respectively, compared to a purely isometric contraction at the same final muscle length and level of activation. This is referred to as the history dependence of force. When matching a given torque level, there is greater muscle activation (electromyography; EMG) following shortening and less activation following lengthening. Owing to these differences in neuromuscular activation, it is unclear whether perception of effort is altered by the history dependence of force during plantar-flexion.MethodsExperiment 1 tested whether perception of effort differed between the rTD and rTE state when torque was matched. Experiment 2 tested whether perception of effort differed between the rTD and rTE state when EMG was matched. Finally, experiment 3 tested whether EMG differed between the rTD and rTE state when perception of effort was matched.ResultsWhen torque was matched, both EMG and perception of effort were higher in the rTD compared to rTE state. When EMG was matched, torque was lower in the rTD compared to rTE state while perception of effort did not differ between the two states. When perception of effort was matched, torque was lower in the rTD compared to rTE state and EMG did not differ between the two states.ConclusionThe combined results from these experiments indicate that the history dependence of force alters one’s perception of effort, dependent on the level of motor command.
Journal Article
The effects of caffeine and d-amphetamine on spatial span task in healthy participants
Studies that examined the effect of amphetamine or caffeine on spatial working memory (SWM) and verbal working memory (VWM) have used various tasks. However, there are no studies that have used spatial span tasks (SSTs) to assess the SWM effect of amphetamine and caffeine, although some studies have used digit span tasks (DST) to assess VWM. Previous reports also showed that increasing dopamine increases psychosis-like experiences (PLE, or schizotypy) scores which are in turn negatively associated with WM performance in people with high schizotypy and people with schizophrenia. Therefore, the present study aimed to examine the influence of d-amphetamine (0.45 mg/kg, PO), a dopamine releasing stimulant, on SST, DST, and on PLE in healthy volunteers. In a separate study, we examined the effect of caffeine, a nonspecific adenosine receptor antagonist with stimulant properties, on similar tasks. Methods: Healthy participants (N = 40) took part in two randomized, double-blind, counter-balanced placebo-controlled cross-over pilot studies: The first group (N = 20) with d-amphetamine (0.45 mg/kg, PO) and the second group (N = 20) with caffeine (200 mg, PO). Spatial span and digit span were examined under four delay conditions (0, 2, 4, 8 s). PLE were assessed using several scales measuring various aspects of psychosis and schizotypy. Results: We failed to find an effect of d-amphetamine or caffeine on SWM or VWM, relative to placebo. However, d-amphetamine increased a composite score of psychosis-like experiences (p = 0.0005), specifically: Scores on Brief Psychiatric Rating Scale, Perceptual Aberrations Scale, and Magical Ideation Scale were increased following d-amphetamine. The degree of change in PLE following d-amphetamine negatively and significantly correlated with changes in SWM, mainly at the longest delay condition of 8 s (r = -0.58, p = 0.006). Conclusion: The present results showed that moderate-high dose of d-amphetamine and moderate dose of caffeine do not directly affect performances on DST or SST. However, the results indicate that d-amphetamine indirectly influences SWM, through its effect on psychosis-like experiences. Trial registration. Clinical Trial Registration Number: CT-2018-CTN-02561 (Therapeutic Goods Administration Clinical Trial Registry) and ACTRN12618001292268 (The Australian New Zealand Clinical Trials Registry) for caffeine study, and ACTRN12608000610336 for d-amphetamine study.
Journal Article
Improved correspondence of fMRI visual field localizer data after cortex-based macroanatomical alignment
Studying the visual system with fMRI often requires using localizer paradigms to define regions of interest (ROIs). However, the considerable interindividual variability of the cerebral cortex represents a crucial confound for group-level analyses. Cortex-based alignment (CBA) techniques reliably reduce interindividual macroanatomical variability. Yet, their utility has not been assessed for visual field localizer paradigms, which map specific parts of the visual field within retinotopically organized visual areas. We evaluated CBA for an attention-enhanced visual field localizer, mapping homologous parts of each visual quadrant in 50 participants. We compared CBA with volume-based alignment and a surface-based analysis, which did not include macroanatomical alignment. CBA led to the strongest increase in the probability of activation overlap (up to 86%). At the group level, CBA led to the most consistent increase in ROI size while preserving vertical ROI symmetry. Overall, our results indicate that in addition to the increased signal-to-noise ratio of a surface-based analysis, macroanatomical alignment considerably improves statistical power. These findings confirm and extend the utility of CBA for the study of the visual system in the context of group analyses. CBA should be particularly relevant when studying neuropsychiatric disorders with abnormally increased interindividual macroanatomical variability.
Journal Article
Prognostic value of texture analysis from cardiac magnetic resonance imaging in patients with Takotsubo syndrome: a machine learning based proof-of-principle approach
Cardiac magnetic resonance (CMR) imaging has become an important technique for non-invasive diagnosis of takotsubo syndrome (TTS). The long-term prognostic value of CMR imaging in TTS has not been fully elucidated yet. This study sought to evaluate the prognostic value of texture analysis (TA) based on CMR images in patients with TTS using machine learning. In this multicenter study (InterTAK Registry), we investigated CMR imaging data of 58 patients (56 women, mean age 68 ± 12 years) with TTS. CMR imaging was performed in the acute to subacute phase (median time after symptom onset 4 days) of TTS. TA of the left ventricle was performed using free-hand regions-of-interest in short axis late gadolinium-enhanced and on T2-weighted (T2w) images. A total of 608 TA features adding the parameters age, gender, and body mass index were included. Dimension reduction was performed removing TA features with poor intra-class correlation coefficients (ICC ≤ 0.6) and those being redundant (correlation matrix with Pearson correlation coefficient r > 0.8). Five common machine-learning classifiers (artificial neural network Multilayer Perceptron, decision tree J48, NaïveBayes, RandomForest, and Sequential Minimal Optimization) with tenfold cross-validation were applied to assess 5-year outcome including major adverse cardiac and cerebrovascular events (MACCE). Dimension reduction yielded 10 TA features carrying prognostic information, which were all based on T2w images. The NaïveBayes machine learning classifier showed overall best performance with a sensitivity of 82.9% (confidence interval (CI) 80–86.2), specificity of 83.7% (CI 75.7–92), and an area-under-the receiver operating characteristics curve of 0.88 (CI 0.83–0.92). This proof-of-principle study is the first to identify unique T2w-derived TA features that predict long-term outcome in patients with TTS. These features might serve as imaging prognostic biomarkers in TTS patients.
Journal Article
Discovery of a cis-regulatory element SaeM involved in dynamic regulation of synergid-specific MYB98
MYB98 is a key regulator of the genetic network behind pollen tube attraction toward the female gametophyte. MYB98 is specifically expressed in the synergid cells (SCs), a female gametophyte component cells specialized for pollen tube attraction. However, it had not been clear how exactly MYB98 achieves this specific expression pattern. In the current study, we have determined that a normal SC-specific expression of MYB98 is dependent on a 16-bp-long cis -regulatory element, CATTTACACATTAAAA, freshly named as the “ S ynergid-specific A ctivation E lement of M YB98 ” ( SaeM ). An 84 bp fragment harboring SaeM in the middle was sufficient to drive exclusively SC-specific expression. The element was present in a significantly large proportion of SC-specific gene promoters and in the promoter of MYB98 homologous genes in the Brassicaceae ( pMYB98 s). Significance of such family-wide SaeM-like element conservation in exclusive SC-specific expression was confirmed by the Arabidopsis-like activation feature of Brassica oleracea– derived pMYB98 and absence of such feature of pMYB98 derived from a non-Brassicaceae member Prunus persica . Additionally, the yeast-one-hybrid assay showed that the SaeM can be recognized by ANTHOCYANINLESS2 (ANL2) and DAP-seq data further suggested for additional three ANL2 homologs targeting the similar cis- element. Overall, our study has concluded that SaeM plays a crucial role in driving exclusively SC-specific expression of MYB98 and strongly suggests for the involvement of ANL2 and its homologs in its dynamic regulation in planta . Future study on the transcription factors is expected to shed more light on the mechanism behind the process.
Journal Article