1
|
Tik N, Gal S, Madar A, Ben-David T, Bernstein-Eliav M, Tavor I. Generalizing prediction of task-evoked brain activity across datasets and populations. Neuroimage 2023; 276:120213. [PMID: 37268097 DOI: 10.1016/j.neuroimage.2023.120213] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/27/2023] [Revised: 05/28/2023] [Accepted: 05/30/2023] [Indexed: 06/04/2023] Open
Abstract
Predictions of task-based functional magnetic resonance imaging (fMRI) from task-free resting-state (rs) fMRI have gained popularity over the past decade. This method holds a great promise for studying individual variability in brain function without the need to perform highly demanding tasks. However, in order to be broadly used, prediction models must prove to generalize beyond the dataset they were trained on. In this work, we test the generalizability of prediction of task-fMRI from rs-fMRI across sites, MRI vendors and age-groups. Moreover, we investigate the data requirements for successful prediction. We use the Human Connectome Project (HCP) dataset to explore how different combinations of training sample sizes and number of fMRI datapoints affect prediction success in various cognitive tasks. We then apply models trained on HCP data to predict brain activations in data from a different site, a different MRI vendor (Phillips vs. Siemens scanners) and a different age group (children from the HCP-development project). We demonstrate that, depending on the task, a training set of approximately 20 participants with 100 fMRI timepoints each yields the largest gain in model performance. Nevertheless, further increasing sample size and number of timepoints results in significantly improved predictions, until reaching approximately 450-600 training participants and 800-1000 timepoints. Overall, the number of fMRI timepoints influences prediction success more than the sample size. We further show that models trained on adequate amounts of data successfully generalize across sites, vendors and age groups and provide predictions that are both accurate and individual-specific. These findings suggest that large-scale publicly available datasets may be utilized to study brain function in smaller, unique samples.
Collapse
Affiliation(s)
- Niv Tik
- Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel
| | - Shachar Gal
- Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel
| | - Asaf Madar
- Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel
| | - Tamar Ben-David
- Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel
| | - Michal Bernstein-Eliav
- Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel
| | - Ido Tavor
- Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel; Strauss Center for Computational Neuroimaging, Tel Aviv University, Tel Aviv, Israel.
| |
Collapse
|
2
|
Rackes A, Ben-David T, Waring MS. Outcome-based ventilation: A framework for assessing performance, health, and energy impacts to inform office building ventilation decisions. Indoor Air 2018; 28:585-603. [PMID: 29683212 DOI: 10.1111/ina.12466] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 08/04/2017] [Accepted: 04/12/2018] [Indexed: 06/08/2023]
Abstract
This article presents an outcome-based ventilation (OBV) framework, which combines competing ventilation impacts into a monetized loss function ($/occ/h) used to inform ventilation rate decisions. The OBV framework, developed for U.S. offices, considers six outcomes of increasing ventilation: profitable outcomes realized from improvements in occupant work performance and sick leave absenteeism; health outcomes from occupant exposure to outdoor fine particles and ozone; and energy outcomes from electricity and natural gas usage. We used the literature to set low, medium, and high reference values for OBV loss function parameters, and evaluated the framework and outcome-based ventilation rates using a simulated U.S. office stock dataset and a case study in New York City. With parameters for all outcomes set at medium values derived from literature-based central estimates, higher ventilation rates' profitable benefits dominated negative health and energy impacts, and the OBV framework suggested ventilation should be ≥45 L/s/occ, much higher than the baseline ~8.5 L/s/occ rate prescribed by ASHRAE 62.1. Only when combining very low parameter estimates for profitable impacts with very high ones for health and energy impacts were all outcomes on the same order. Even then, however, outcome-based ventilation rates were often twice the baseline rate or more.
Collapse
Affiliation(s)
- A Rackes
- Department of Civil, Architectural and Environmental Engineering, Drexel University, Philadelphia, PA, USA
| | - T Ben-David
- Department of Civil, Architectural and Environmental Engineering, Drexel University, Philadelphia, PA, USA
| | - M S Waring
- Department of Civil, Architectural and Environmental Engineering, Drexel University, Philadelphia, PA, USA
| |
Collapse
|