RESUMO
A growing number of studies apply deep neural networks (DNNs) to recordings of human electroencephalography (EEG) to identify a range of disorders. In many studies, EEG recordings are split into segments, and each segment is randomly assigned to the training or test set. As a consequence, data from individual subjects appears in both the training and the test set. Could high test-set accuracy reflect data leakage from subject-specific patterns in the data, rather than patterns that identify a disease? We address this question by testing the performance of DNN classifiers using segment-based holdout (in which segments from one subject can appear in both the training and test set), and comparing this to their performance using subject-based holdout (where all segments from one subject appear exclusively in either the training set or the test set). In two datasets (one classifying Alzheimer's disease, and the other classifying epileptic seizures), we find that performance on previously-unseen subjects is strongly overestimated when models are trained using segment-based holdout. Finally, we survey the literature and find that the majority of translational DNN-EEG studies use segment-based holdout. Most published DNN-EEG studies may dramatically overestimate their classification performance on new subjects.
RESUMO
The neural and perceptual effects of attention were traditionally assumed to be sustained over time, but recent work suggests that covert attention rhythmically switches between objects at 3-8 Hz. Here I use simulations to demonstrate that the analysis approaches commonly used to test for rhythmic oscillations generate false positives in the presence of aperiodic temporal structure. I then propose two alternative analyses that are better able to discriminate between periodic and aperiodic structure in time series. Finally, I apply these alternative analyses to published datasets and find no evidence for behavioural rhythms in attentional switching after accounting for aperiodic temporal structure. The techniques presented here will help clarify the periodic and aperiodic dynamics of perception and of cognition more broadly.
Assuntos
Atenção , Percepção Visual , Cognição , Humanos , Estimulação Luminosa/métodos , Fatores de TempoRESUMO
How does the brain anticipate information in language? When people perceive speech, low-frequency (<10 Hz) activity in the brain synchronizes with bursts of sound and visual motion. This phenomenon, called cortical stimulus-tracking, is thought to be one way that the brain predicts the timing of upcoming words, phrases, and syllables. In this study, we test whether stimulus-tracking depends on domain-general expertise or on language-specific prediction mechanisms. We go on to examine how the effects of expertise differ between frontal and sensory cortex. We recorded electroencephalography (EEG) from human participants who were experts in either sign language or ballet, and we compared stimulus-tracking between groups while participants watched videos of sign language or ballet. We measured stimulus-tracking by computing coherence between EEG recordings and visual motion in the videos. Results showed that stimulus-tracking depends on domain-general expertise, and not on language-specific prediction mechanisms. At frontal channels, fluent signers showed stronger coherence to sign language than to dance, whereas expert dancers showed stronger coherence to dance than to sign language. At occipital channels, however, the two groups of participants did not show different patterns of coherence. These results are difficult to explain by entrainment of endogenous oscillations, because neither sign language nor dance show any periodicity at the frequencies of significant expertise-dependent stimulus-tracking. These results suggest that the brain may rely on domain-general predictive mechanisms to optimize perception of temporally-predictable stimuli such as speech, sign language, and dance.
Assuntos
Eletroencefalografia , Fala , Atenção , Encéfalo , Humanos , PeriodicidadeRESUMO
How do humans compute approximate number? According to one influential theory, approximate number representations arise in the intraparietal sulcus and are amodal, meaning that they arise independent of any sensory modality. Alternatively, approximate number may be computed initially within sensory systems. Here we tested for sensitivity to approximate number in the visual system using steady state visual evoked potentials. We recorded electroencephalography from humans while they viewed dotclouds presented at 30 Hz, which alternated in numerosity (ranging from 10 to 20 dots) at 15 Hz. At this rate, each dotcloud backward masked the previous dotcloud, disrupting top-down feedback to visual cortex and preventing conscious awareness of the dotclouds' numerosities. Spectral amplitude at 15 Hz measured over the occipital lobe (Oz) correlated positively with the numerical ratio of the stimuli, even when nonnumerical stimulus attributes were controlled, indicating that subjects' visual systems were differentiating dotclouds on the basis of their numerical ratios. Crucially, subjects were unable to discriminate the numerosities of the dotclouds consciously, indicating the backward masking of the stimuli disrupted reentrant feedback to visual cortex. Approximate number appears to be computed within the visual system, independently of higher-order areas, such as the intraparietal sulcus.
Assuntos
Potenciais Evocados Visuais/fisiologia , Conceitos Matemáticos , Córtex Visual/fisiologia , Adulto , Estado de Consciência/fisiologia , Eletroencefalografia , Feminino , Humanos , Masculino , Estimulação Luminosa , Percepção Visual/fisiologiaRESUMO
Different regions of the human cerebral cortex are specialized for different emotions, but the principles underlying this specialization have remained unknown. According to the sword and shield hypothesis, hemispheric specialization for affective motivation, a basic dimension of human emotion, varies across individuals according to the way they use their hands to perform approach- and avoidance-related actions. In a test of this hypothesis, here we measured approach motivation before and after five sessions of transcranial direct current stimulation to increase excitation in the left or right dorsolateral prefrontal cortex, in healthy adults whose handedness ranged from strongly left-handed to strongly right-handed. The strength and direction of participants' handedness predicted whether electrical stimulation to frontal cortex caused an increase or decrease in their experience of approach-related emotions. The organization of approach motivation in the human cerebral cortex varies across individuals as predicted by the organization of the individuals' motor systems. These results show that the large-scale cortical organization of abstract concepts corresponds with the way people use their hands to interact with the world. Affective motivation may re-use neural circuits that evolved for performing approach- and avoidance-related motor actions.This article is part of the theme issue 'Varieties of abstract concepts: development, use and representation in the brain'.
Assuntos
Emoções/fisiologia , Lateralidade Funcional , Motivação/fisiologia , Córtex Pré-Frontal/fisiologia , Humanos , Estimulação Transcraniana por Corrente ContínuaRESUMO
Despite immense variability across languages, people can learn to understand any human language, spoken or signed. What neural mechanisms allow people to comprehend language across sensory modalities? When people listen to speech, electrophysiological oscillations in auditory cortex entrain to slow ([Formula: see text]8 Hz) fluctuations in the acoustic envelope. Entrainment to the speech envelope may reflect mechanisms specialized for auditory perception. Alternatively, flexible entrainment may be a general-purpose cortical mechanism that optimizes sensitivity to rhythmic information regardless of modality. Here, we test these proposals by examining cortical coherence to visual information in sign language. First, we develop a metric to quantify visual change over time. We find quasiperiodic fluctuations in sign language, characterized by lower frequencies than fluctuations in speech. Next, we test for entrainment of neural oscillations to visual change in sign language, using electroencephalography (EEG) in fluent speakers of American Sign Language (ASL) as they watch videos in ASL. We find significant cortical entrainment to visual oscillations in sign language <5 Hz, peaking at [Formula: see text]1 Hz. Coherence to sign is strongest over occipital and parietal cortex, in contrast to speech, where coherence is strongest over the auditory cortex. Nonsigners also show coherence to sign language, but entrainment at frontal sites is reduced relative to fluent signers. These results demonstrate that flexible cortical entrainment to language does not depend on neural processes that are specific to auditory speech perception. Low-frequency oscillatory entrainment may reflect a general cortical mechanism that maximizes sensitivity to informational peaks in time-varying signals.
Assuntos
Língua de Sinais , Córtex Visual/fisiologia , Adulto , Córtex Cerebral/fisiologia , Eletroencefalografia , Fenômenos Eletrofisiológicos , Feminino , Humanos , Testes de Linguagem , Masculino , Estimulação Luminosa , Gravação em Vídeo , Percepção Visual/fisiologia , Adulto JovemRESUMO
Neuroimaging and brain damage studies suggest that dorsolateral prefrontal cortex (dlPFC) is involved in the cognitive control of episodic recollection. If dlPFC is causally involved in retrieval, then transcranial direct current stimulation (tDCS) of this brain region should increase recollection accuracy, especially when recollection is difficult and requires cognitive control. Here, we report the first brain stimulation experiment to directly test this hypothesis. We administered tDCS to dlPFC immediately after studying to-be-learned material but just prior to recollection testing, thereby targeting retrieval processes. We found that stimulation of dlPFC significantly increased recollection accuracy, relative to a no-stimulation sham condition and also relative to active stimulation of a comparison region in left parietal cortex. There was no significant difference in the size of this increase between hemispheres. Moreover, these dlPFC stimulation effects were behaviorally selective, increasing accuracy only when participants needed to recollect difficult information. Electrically stimulating dlPFC allowed people to more accurately recollect specific details of their experiences, demonstrating a causal role of dlPFC in the retrieval of episodic memories.
Assuntos
Memória/fisiologia , Rememoração Mental/fisiologia , Córtex Pré-Frontal/fisiologia , Tempo de Reação/fisiologia , Estimulação Transcraniana por Corrente Contínua , Adolescente , Adulto , Feminino , Humanos , Testes Neuropsicológicos , Adulto JovemRESUMO
In two experiments, Brookshire, Ivry, and Casasanto (2010) showed that words with positive and negative emotional valence can activate spatial representations with a high degree of automaticity, but also that this activation is highly context dependent. Lebois, Wilson-Mendenhall, and Barsalou (2015) reported that they "aimed to replicate" our study but found only null results in the "Brookshire et al. replication" conditions. Here we express concerns about three aspects of this paper. First, the study was not an attempt to replicate ours; it was a different study that adapted our method. Second, Lebois et al. did not accurately represent our theoretical position. Third, Lebois et al.'s main conclusion, that spatial congruity effects depend on the task context, was not supported by their data. Despite these concerns, we agree with Lebois et al.'s overall message that spatial aspects of words' meanings are activated differently in different contexts. This was a main conclusion of our study as well.
Assuntos
Emoções , Semântica , Humanos , ReflexoRESUMO
BACKGROUND: According to decades of research on affective motivation in the human brain, approach motivational states are supported primarily by the left hemisphere and avoidance states by the right hemisphere. The underlying cause of this specialization, however, has remained unknown. Here we conducted a first test of the Sword and Shield Hypothesis (SSH), according to which the hemispheric laterality of affective motivation depends on the laterality of motor control for the dominant hand (i.e., the "sword hand," used preferentially to perform approach actions) and the nondominant hand (i.e., the "shield hand," used preferentially to perform avoidance actions). METHODOLOGY/PRINCIPAL FINDINGS: To determine whether the laterality of approach motivation varies with handedness, we measured alpha-band power (an inverse index of neural activity) in right- and left-handers during resting-state electroencephalography and analyzed hemispheric alpha-power asymmetries as a function of the participants' trait approach motivational tendencies. Stronger approach motivation was associated with more left-hemisphere activity in right-handers, but with more right-hemisphere activity in left-handers. CONCLUSIONS: The hemispheric correlates of approach motivation reversed between right- and left-handers, consistent with the way they typically use their dominant and nondominant hands to perform approach and avoidance actions. In both right- and left-handers, approach motivation was lateralized to the same hemisphere that controls the dominant hand. This covariation between neural systems for action and emotion provides initial support for the SSH.