Gon García-Castro
@gongcastro.bsky.social
Clueless postdoc investigating language acquisition and comparative cognition • fNIRS and Bayesian stats • él/he
🌐 gongcastro.github.io
📍Institut de Recerca Sant Joan de Déu, previously at Universitat Pompeu Fabra.
🌐 gongcastro.github.io
📍Institut de Recerca Sant Joan de Déu, previously at Universitat Pompeu Fabra.
Tidal también tiene wrapped, y me complace anunciar que, a favor de todo pronóstico, Amaral sigue ahí.
December 4, 2025 at 8:25 AM
Tidal también tiene wrapped, y me complace anunciar que, a favor de todo pronóstico, Amaral sigue ahí.
I illustrate some use cases, modifying the duration of individual phonemes or modulating their pitch contours. Here's a simple example.
November 8, 2025 at 1:45 PM
I illustrate some use cases, modifying the duration of individual phonemes or modulating their pitch contours. Here's a simple example.
Made some friends in the Marmara sea.
October 21, 2025 at 6:40 AM
Made some friends in the Marmara sea.
Barcelona 🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸
October 4, 2025 at 11:48 AM
Barcelona 🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸
This week I taught an introductory lecture on fNIRS at the Deep Tech Brain Summer School at @ub.edu.
We covered some theory, experimental design in dev populations and a live demonstration of data acquisition and preprocessing using @nirx-nirs.bsky.social's NIRSport2. Thanks to the organisers!
We covered some theory, experimental design in dev populations and a live demonstration of data acquisition and preprocessing using @nirx-nirs.bsky.social's NIRSport2. Thanks to the organisers!
July 23, 2025 at 7:17 AM
This week I taught an introductory lecture on fNIRS at the Deep Tech Brain Summer School at @ub.edu.
We covered some theory, experimental design in dev populations and a live demonstration of data acquisition and preprocessing using @nirx-nirs.bsky.social's NIRSport2. Thanks to the organisers!
We covered some theory, experimental design in dev populations and a live demonstration of data acquisition and preprocessing using @nirx-nirs.bsky.social's NIRSport2. Thanks to the organisers!
There's some audacity in this statement after a decade of promoting the mantra that there is a "replicability *crisis*" that should make everyone doubt the credibility of science, as a whole.
Now it turns out it's been fueling anti-science discourses without improving the credibility of science?
Now it turns out it's been fueling anti-science discourses without improving the credibility of science?
June 29, 2025 at 10:07 AM
There's some audacity in this statement after a decade of promoting the mantra that there is a "replicability *crisis*" that should make everyone doubt the credibility of science, as a whole.
Now it turns out it's been fueling anti-science discourses without improving the credibility of science?
Now it turns out it's been fueling anti-science discourses without improving the credibility of science?
Back with a late update from #WILD2025 in Donosti!
Flora Chartier (our new PhD star ✨) and I presented exciting, preliminary results from our studies. Thanks everyone for the feedback!
Here's a glance of our poster on the ongoing fNIRS study in newborns at the @sjdhospitalbarcelona.org.
Flora Chartier (our new PhD star ✨) and I presented exciting, preliminary results from our studies. Thanks everyone for the feedback!
Here's a glance of our poster on the ongoing fNIRS study in newborns at the @sjdhospitalbarcelona.org.
June 20, 2025 at 7:54 AM
Back with a late update from #WILD2025 in Donosti!
Flora Chartier (our new PhD star ✨) and I presented exciting, preliminary results from our studies. Thanks everyone for the feedback!
Here's a glance of our poster on the ongoing fNIRS study in newborns at the @sjdhospitalbarcelona.org.
Flora Chartier (our new PhD star ✨) and I presented exciting, preliminary results from our studies. Thanks everyone for the feedback!
Here's a glance of our poster on the ongoing fNIRS study in newborns at the @sjdhospitalbarcelona.org.
Ens veiem el dissabte a Barcelona. 🇵🇸🇵🇸🇱🇧🇱🇧
June 12, 2025 at 8:03 AM
Ens veiem el dissabte a Barcelona. 🇵🇸🇵🇸🇱🇧🇱🇧
Being caught between supporting a fascist dictator and not supporting a fascist dictator is now called "it's complicated".
March 26, 2025 at 5:43 PM
Being caught between supporting a fascist dictator and not supporting a fascist dictator is now called "it's complicated".
In Exp. 3, we replicated Exp. 1, but this time we asked participants to report, for each Catalan or Spanish word presented, whether they were familiar with its meaning. We excluded these responses from the data analysis. Results were equivalent to those in Exp. 1.
March 25, 2025 at 1:55 PM
In Exp. 3, we replicated Exp. 1, but this time we asked participants to report, for each Catalan or Spanish word presented, whether they were familiar with its meaning. We excluded these responses from the data analysis. Results were equivalent to those in Exp. 1.
Of course, some participants might have been slightly pessimistic about their Catalan or Spanish knowledge, being familiar with some of the (high frequency) Catalan or Spanish words presented, and their meanings.
This was the case for some words, with some surprising examples!
This was the case for some words, with some surprising examples!
March 25, 2025 at 1:55 PM
Of course, some participants might have been slightly pessimistic about their Catalan or Spanish knowledge, being familiar with some of the (high frequency) Catalan or Spanish words presented, and their meanings.
This was the case for some words, with some surprising examples!
This was the case for some words, with some surprising examples!
In Exp. 2, we tested Spanish natives in the same task, this time listening to only Catalan (typologically closer to Spanish than English). Again, participants reported no knowledge of Catalan. Spanish natives also did quite well and performed slightly better than English natives.
March 25, 2025 at 1:55 PM
In Exp. 2, we tested Spanish natives in the same task, this time listening to only Catalan (typologically closer to Spanish than English). Again, participants reported no knowledge of Catalan. Spanish natives also did quite well and performed slightly better than English natives.
Provided at least some phonological similarity between the presented Catalan or Spanish words and their correct English translations (and few higher-frequency phonological neighbours), participants were surprisingly good at translating Spanish!
March 25, 2025 at 1:55 PM
Provided at least some phonological similarity between the presented Catalan or Spanish words and their correct English translations (and few higher-frequency phonological neighbours), participants were surprisingly good at translating Spanish!
In Exp. 1, English natives who reported no prior knowledge of Catalan or Spanish (or any other Romance language), completed a translation elicitation task.
Participants listened to a series of Catalan or Spanish words, and typed their best-guess English translation of the word.
Participants listened to a series of Catalan or Spanish words, and typed their best-guess English translation of the word.
March 25, 2025 at 1:55 PM
In Exp. 1, English natives who reported no prior knowledge of Catalan or Spanish (or any other Romance language), completed a translation elicitation task.
Participants listened to a series of Catalan or Spanish words, and typed their best-guess English translation of the word.
Participants listened to a series of Catalan or Spanish words, and typed their best-guess English translation of the word.
New preprint w/Serene Siow, Kim Plunkett, and Nuria Sebastian-Galles.
We investigated how listeners exploit phonological similarity to correctly translate words from a completely unfamiliar language!🧵
🔗 osf.io/7k3ga_v1
@upf.edu @ox.ac.uk
We investigated how listeners exploit phonological similarity to correctly translate words from a completely unfamiliar language!🧵
🔗 osf.io/7k3ga_v1
@upf.edu @ox.ac.uk
March 25, 2025 at 1:55 PM
New preprint w/Serene Siow, Kim Plunkett, and Nuria Sebastian-Galles.
We investigated how listeners exploit phonological similarity to correctly translate words from a completely unfamiliar language!🧵
🔗 osf.io/7k3ga_v1
@upf.edu @ox.ac.uk
We investigated how listeners exploit phonological similarity to correctly translate words from a completely unfamiliar language!🧵
🔗 osf.io/7k3ga_v1
@upf.edu @ox.ac.uk
Barcelona 🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸
January 18, 2025 at 11:56 AM
Barcelona 🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸🇵🇸
WE'RE SO BACK
(made a flask app for setting up the new babylab, added unit and functional tests, it does the job, I hope to finally write about it soon.)
(made a flask app for setting up the new babylab, added unit and functional tests, it does the job, I hope to finally write about it soon.)
January 13, 2025 at 2:47 PM
WE'RE SO BACK
(made a flask app for setting up the new babylab, added unit and functional tests, it does the job, I hope to finally write about it soon.)
(made a flask app for setting up the new babylab, added unit and functional tests, it does the job, I hope to finally write about it soon.)
Acabo de terminar «La Yihad Butleriana». Apenas recuerdo engancharme tan rápido y fuerte a un libro.
Después de casi tener que abandonar «Mesias de Dune» de aburrimiento no os imagináis lo fácil que ha sido acabarme este. No es de los originales y sin embargo me parece el mejor de la saga. #Dune
Después de casi tener que abandonar «Mesias de Dune» de aburrimiento no os imagináis lo fácil que ha sido acabarme este. No es de los originales y sin embargo me parece el mejor de la saga. #Dune
January 1, 2025 at 10:59 AM
Acabo de terminar «La Yihad Butleriana». Apenas recuerdo engancharme tan rápido y fuerte a un libro.
Después de casi tener que abandonar «Mesias de Dune» de aburrimiento no os imagináis lo fácil que ha sido acabarme este. No es de los originales y sin embargo me parece el mejor de la saga. #Dune
Después de casi tener que abandonar «Mesias de Dune» de aburrimiento no os imagináis lo fácil que ha sido acabarme este. No es de los originales y sin embargo me parece el mejor de la saga. #Dune
Feliz Navidad!! 🌟
December 26, 2024 at 4:10 PM
Feliz Navidad!! 🌟
I've been relatively inactive on social media lately due to a bunch of (mostly good) recent developments. Overall, I'm happiest than even, moisturised, and in my lane. Writing up some blog posts on recent work I've done setting up the more technical aspects of our new group, which has been exciting!
December 15, 2024 at 10:52 AM
I've been relatively inactive on social media lately due to a bunch of (mostly good) recent developments. Overall, I'm happiest than even, moisturised, and in my lane. Writing up some blog posts on recent work I've done setting up the more technical aspects of our new group, which has been exciting!
Here's a very small contribution to the #fNIRS community that has welcomed me so generously!
I made a simple #Python function to calculate the DPF following the Scholkmann method. Here's a post in which I reproduce the figures in the original paper:
🔗 gongcastro.github.io/blog/dpf-sch...
I made a simple #Python function to calculate the DPF following the Scholkmann method. Here's a post in which I reproduce the figures in the original paper:
🔗 gongcastro.github.io/blog/dpf-sch...
November 19, 2024 at 7:23 AM
Here's a very small contribution to the #fNIRS community that has welcomed me so generously!
I made a simple #Python function to calculate the DPF following the Scholkmann method. Here's a post in which I reproduce the figures in the original paper:
🔗 gongcastro.github.io/blog/dpf-sch...
I made a simple #Python function to calculate the DPF following the Scholkmann method. Here's a post in which I reproduce the figures in the original paper:
🔗 gongcastro.github.io/blog/dpf-sch...
Even as models of cognition, LLMs are pretty bad. As the model upscales to approximate human-like outputs, it loses its usefulness as an abstract, interpretable simplification of cognition.
Jorge Luis Borges wrote a beautiful story about this:
Jorge Luis Borges wrote a beautiful story about this:
November 18, 2024 at 8:12 AM
Even as models of cognition, LLMs are pretty bad. As the model upscales to approximate human-like outputs, it loses its usefulness as an abstract, interpretable simplification of cognition.
Jorge Luis Borges wrote a beautiful story about this:
Jorge Luis Borges wrote a beautiful story about this:
3) This co-activation results in an asymmetrical accumulation of learning instances across languages: words in the lower-exposure language benefit more often from the "additional activation" of their translation equivalents, to which infants are exposed more frequently. 📈
November 16, 2024 at 9:41 PM
3) This co-activation results in an asymmetrical accumulation of learning instances across languages: words in the lower-exposure language benefit more often from the "additional activation" of their translation equivalents, to which infants are exposed more frequently. 📈
The 1st paper of my thesis out in Child Development, OA! 🎉
«Cognate beginnings to bilingual lexical acquisition»
w/Daniela Avila-Varela, Ignacio Castillejo & Nuria Sebastian-Galles.
srcd.onlinelibrary.wiley.com/doi/full/10....
«Cognate beginnings to bilingual lexical acquisition»
w/Daniela Avila-Varela, Ignacio Castillejo & Nuria Sebastian-Galles.
srcd.onlinelibrary.wiley.com/doi/full/10....
November 16, 2024 at 9:41 PM
The 1st paper of my thesis out in Child Development, OA! 🎉
«Cognate beginnings to bilingual lexical acquisition»
w/Daniela Avila-Varela, Ignacio Castillejo & Nuria Sebastian-Galles.
srcd.onlinelibrary.wiley.com/doi/full/10....
«Cognate beginnings to bilingual lexical acquisition»
w/Daniela Avila-Varela, Ignacio Castillejo & Nuria Sebastian-Galles.
srcd.onlinelibrary.wiley.com/doi/full/10....