Bambi: A simple interface for fitting Bayesian linear models in Python

Capretto, Tomás; Piho, Camen; Kumar, Ravin; Westfall, Jacob; Yarkoni, Tal; Martin, Osvaldo A.

doi:10.48550/arxiv.2012.10754

Cited by 9 publications

(6 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To investigate the influence of individual prediction tendencies on speech tracking under different noise conditions, we used Bayesian multilevel regression models with Bambi (Capretto et al, 2022), a python package built on top of the PyMC3 package (Salvatier et al, 2016) for probabilistic programming. The correlation between predicted brain activity from speech envelope encoding and true brain activity was used as dependent variable, and separate models were calculated per voxel using the following formula according to the Wilkinson notation (Wilkinson & Rogers, 1973): To investigate the influence of higher-level probabilistic structure of speech, we also calculated a model for which the dependent variable only included cortical tracking (i. e. speech envelope encoding) results of lexically identical nouns of high vs. low surprisal: Before entering the models, prediction_tendency was zero-centered (note that in these interaction models other predictors are assumed to be zero to estimate effects of one predictor) and the number of distractors (0-2) was treated as a continuous variable.…”

Section: Methodsmentioning

confidence: 99%

“…The correlation between predicted brain activity from speech envelope encoding and true brain activity was used as dependent variable, and separate models were calculated per voxel using the following formula according to the Wilkinson notation (Wilkinson & Rogers, 1973): To investigate the influence of higher-level probabilistic structure of speech, we also calculated a model for which the dependent variable only included cortical tracking (i. e. speech envelope encoding) results of lexically identical nouns of high vs. low surprisal: Before entering the models, prediction_tendency was zero-centered (note that in these interaction models other predictors are assumed to be zero to estimate effects of one predictor) and the number of distractors (0-2) was treated as a continuous variable. As priors we used the weakly-or non-informative default priors of Bambi (Capretto et al, 2022). For a summary of model parameters we report regression coefficients and 94% high density intervals (HDI) of the posterior distribution.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Individual prediction tendencies facilitate cortical speech tracking

Schubert

Schmidt

Gehmacher

et al. 2022

Preprint

View full text Add to dashboard Cite

Listening can be conceptualized as a process of active inference, in which the brain forms internal models to predict and integrate auditory information in a complex interaction of bottom-up and top-down processes. Whether inter-individual “prediction tendencies” shape listening experiences of real-world stimuli such as speech is, however, unknown. In the current study, we used a passive paradigm presenting tone sequences of varying entropy level, to independently quantify auditory prediction tendency (as the tendency to anticipate low-level acoustic features according to their contextual probability) for each individual. This measure was then used to predict the magnitude of cortical speech (envelope) tracking in a multi speaker listening task, where participants listened to audiobooks narrated by a target speaker in isolation or interfered by 1 or 2 distractors. Furthermore, rare semantic violations were introduced into the story, enabling us to also examine effects of word surprisal during continuous speech processing. Our results show that individual prediction tendency facilitates cortical speech tracking. Furthermore, we find interactions between individual prediction tendency and background noise as well as word surprisal in disparate brain regions. In sum, our findings suggest that individual prediction tendencies are generalizable across different listening situations and may serve as a valuable element to explain interindividual differences in natural listening experience.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Individual prediction tendencies facilitate cortical speech tracking

Schubert

Schmidt

Gehmacher

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…To investigate the relationship between e.g. age and aperiodic activity recorded using MEG, we used bayesian generalized linear models (GLMs) either build directly in PyMC (a python package for probabilistic programming; (Salvatier et al, 2016)) or in Bambi (a high-level interface to PyMC; (Capretto et al, 2022)). Decisions for either Bambi or PyMC were made based on the accessibility of appropriate statistical families for the respective dependent variables in a GLM.…”

Section: Methodsmentioning

confidence: 99%

Age-related changes in “cortical” 1/f dynamics are linked to cardiac activity

Schmidt

Danböck

Trinka

et al. 2022

Preprint

View full text Add to dashboard Cite

The power of electrophysiologically measured cortical activity decays 1/fX. The slope of this decay (i.e. the spectral exponent) is modulated by various factors such as age, cognitive states or psychiatric/neurological disorders. Interestingly, a mostly parallel line of research has shown similar effects for the spectral exponent in the electrocardiogram (ECG). This raises the question whether these bodywide changes in spectral exponents are independent or if the 'cortically' measured decay in power is explained by cardiac activity captured by cortical sensors (via volume conduction). Focusing on well-established age-related changes in spectral exponents, we analyzed a total of 1282 recordings of magnetoencephalography (MEG) resting state measurements with concurrent ECG in an age-diverse sample. We show that age-related changes of the spectral exponent of neural power spectra are largely explained by cardiac activity. Our results raise concerns when interpreting aperiodic activity as 'cortical' in absence of ECG recordings.

show abstract

“…To estimate this model, we used Bayesian methods, which can flexibly estimate the parameters of non-standard models. Models were estimated using the Bambi (Capretto et al, 2022) and PyMC3 (Salvatier et al, 2016) packages in Python. Before fitting the model, both predictors were z-score transformed, and the dependent variable (morph percentage) was scaled to the interval (0, 1) using the method suggested by Smithson and Verkuilen (2006).…”

Section: Regression Analytic Strategymentioning

confidence: 99%

Incomplete faces are completed using a more average face

Kramer

Jones

2022

Cogn. Research

View full text Add to dashboard Cite

Facial first impressions are known to influence how we behave towards others. As a result of the COVID-19 pandemic, we often view incomplete faces due to the commonplace wearing of face masks. Previous research has shown that perceptions of attractiveness are often increased due to these coverings, with initial evidence suggesting that this may be caused by viewers using a mental representation of the average face to complete any missing information. Here, we directly address this hypothesis by presenting participants with incomplete faces (either the lower or upper half removed) and asking them to decide how they thought the actual, full face looked. Participants were able to manipulate the missing half of the face onscreen by increasing or decreasing the averageness of its shape. Our results demonstrated that participants did not select the original versions of the faces but instead chose more average versions when manipulating both the lower and upper face. Further, the typicality of the original image influenced responses, with less typical faces (in comparison with more typical ones) being completed using an even more average version of the missing half of the faces. Taken together, these findings provide the first direct evidence that people utilise an average/typical internal representation when inferring information about incomplete faces. This result has theoretical importance in terms of visual perception, as well as real-world relevance in a time where face masks are commonplace due to the COVID-19 pandemic.

show abstract

Bambi: A simple interface for fitting Bayesian linear models in Python

Cited by 9 publications

References 18 publications

Individual prediction tendencies facilitate cortical speech tracking

Individual prediction tendencies facilitate cortical speech tracking

Age-related changes in “cortical” 1/f dynamics are linked to cardiac activity

Incomplete faces are completed using a more average face

Contact Info

Product

Resources

About