Brain activity atlas (BAA) aims to elucidate the brain activity across a wide array of domains (e.g., object recognition, motion perception, working memory, attention, language, and theory of mind) using the task-evoked fMRI (tfMRI). The brain activity for a comprehensive battery of functional localizers will be recorded and characterized on multiple levels as follows. First, subject-specific activity atlas (SAA) for the contrasts of interest will be derived for each individual including subject-specific response magnitude maps (SRM) and subject-specific activation maps (SAM). Second, population-probabilistic activity atlas(PAA) for the contrasts of interest will be derived by summarizing the subject-specific activity maps on voxel level including the probabilistic activation map (PAM), the response magnitude maps (RMM), and the response variability maps (RVM). Third, some well-defined functional regions will be identified manually or automatically for each subject, and the probabilistic region maps (PRM) and associated response profiles will be constructed for each region todelineate how the activity varies on the regional level. Fourth, on the network level, the anatomical and functional connectivity profiles of delineated functional regions will be characterized through acquired structural MRI (sMRI), resting-state functional MRI (rfMRI) and diffusion MRI (dMRI) in the same cohort of subjects with the same scanner. Finally, the behavioral significance of the brain activity in voxel, region, and network level will be characterized by relating brain activity to individual differences in cognition, perception, and personality.
Candidate task domains and functional localizer
|Task||Contrast of interest||
Region of interest
|Object recognition(OR)||Category-specific representations localizer with dynamic stimuli||Faces vs. Objects; Scenes vs. objects; Objects vs. Scramble Objects||Category-selective regions, located in occipitotemporal cortex(e.g., OFA, FFA)|
|Motion perception(MP)||Motion detection||Motion vs. static||MT|
|Working memory(WM)||N-back task embedded in category-specific representation task||2-back vs. 1-back||Prefrontal cortex (PFC), inferior frontal gyrus (IFG), precentral gyrus (preCG), anterior cingulate, dorsal parietal.|
|Spatial attention(SA)||Spatial attention task||Peripheral Shifting vs. Central Maintenance||Intraparietal sulcus (IPS), superior parietal lobule (SPL), and middle occipital gyrus (MOG).|
|Feature attention(FA)||rapid serial visual presentation (RSVP)||Non-spatial conjunction vs. feature tasks||Superior parietal lobule (SPL), intraparietal sulcus (IPS), and middle occipital gyrus (MOG).|
|Numerosity(NUM)||Adaptation to Approximate Number||Numerosity change vs. Brightness change||Prefrontal cortex (PFC) and intraparietal sulcus (IPS)|
|Language processing(LAN)||Sentence reading task||Sentences vs. Jabberwocky sentences||interior frontal gyrus(IFG), middle frontal gyrus(MFG),Middle-posterior temporal lobe and angular gyrus.|
|Theory of Mind(ToM)||False belief task||Belief vs. photo stories||temporo-parietal junction(TPJ), precuneus, superior temporal sulcus(STS), and medial prefrontal cortex(mPFC)|
A brain atlas defines spatial and temporal characteristics of the brain such as the location and extent of a given structure or functional region, the response profiles of functional activation, and so on. An atlas can take on many forms, from descriptions of structure or function of individual to maps of groups/ populations. But, technically, an atlas often consists of two parts: the labels and the associated characteristics. The labels provide a reference for the location of region of interest, and the characteristic provides a reference for structural or functional information of the region. Much like geographic atlases can be populated with multiple sources of information about physical, political, economic or environmental aspects of a region of interest, brain atlases may include characteristics about genetic, developmental, evolutionary or physiological or pathological aspects of the brain.
An atlas enables comparison across individuals, modalities or tasks. Differences between species can be catalogued. But in most cases, the value added by brain atlases is the unique and critical ability to integrate information from multiple sources.
- Brain activity atlas
A brain activity atlas defines spatiotemporal characteristic of the brain activity evoked by the specific tasks, especially the extent of the functional regions and associated response profiles.
- Functional localizer
A functional (e.g., fMRI) experiment designed to localize the brain regions which are involved in a specific cognitive process.
- Contrast of interest
A comparison of two or more conditions (e.g., pictures of faces vs. pictures of houses) to reveal the functionally specialized response for a cognitive process (e.g., face perception).
- Response Magnitude
The response intensity for an experimental condition or a contrast of interest, usually measured by the percent signal change (PSC).
- Percent signal change (PSC)
As an index to measure the response magnitude, PSC means the signal change ratio induced by the conditions/contrasts of interest relative to the baseline.
Activation means the statistical significance of the brain activity evoked by the specific task. That, if a voxel show a response that reaches a specified significance level (e.g., Z>2.3), we call it activation.
- Subject-specific activation map (SAM)
The Z- or T-statistic image for the contrast of interest derived from the fixed effect analysis (FFX) of the multiple runs data of an individual.
- Subject-specific response magnitude map (SRM)
SRM refers to the PSC map for the contrast of interest computed from the multiple runs data of an individual.
- Probabilistic Activation map (PAM)
PAM indicates the likelihood of a voxel to be activated for the condition or contrast of interest. It characterizes the spatial variability of activation in population.
- Response magnitude map (RMM)
RMM refers to the PSC map computed from a group of subjects. It characterizes the responses strength in population.
- Response variability map (RVM)
RVM refers to the standard deviation (SD) map of PSC across subjects. It characterizes the responses variability in population.
- Region of interest (ROI)
A set of continuous voxels identified based on the cytoarchitectonic, macroanatomical or functional profiles.
- Functional ROI (fROI)
A ROI identified based on its functional profiles evoked by a functional localizer task (e.g. voxels in the fusiform gyrus that respond selective to faces than other objects).
- Subject-specific ROI (SSR)
The functional ROI identified manually or automatically in a specific subject for the conditions/contrasts of interest.
- Probabilistic fROI map (PRM)
The maps indicate the likelihood that a certain voxel belongs to a specific fROI. It thus quantifies the variability of fROI overlap on voxel level.
- Probabilistic peak map (PPM)
The maps indicate the likelihood that a certain voxel shows peak activation for a specific fROI. It quantifies the spatial variability of peak activation across subjects.
- Maximum probability map of region
The MPM defined the most likely region to which each voxel belonged to and represented all candidate regions in a continuous, but non-overlapping manner in one volume.
- Joint probabilistic activation map (JPAM)
The maps indicate the likelihood of a voxel to be activated in more than one contrast in an individual.
- Conditional probabilistic activation map (CPAM)
The maps quantify the likelihood of a voxel to be activated for a contrast of interest conditional on particular information from other contrasts.
- Marginal probabilistic activation map (MPAM)
The maps give the likelihood of a voxel to be activated for a contrast of interest with no reference to any specific information from other contrasts.
- Joint response magnitude map (JRMM)
The maps describe the likelihood that a voxel takes specific response magnitude on a set of contrasts of interest.
- Conditional response magnitude map (CRMM)
The maps give the likelihood of a voxel takes specific response magnitude for a contrast of interest conditional on particular information from other contrasts.
- Marginal response magnitude map (MRMM)
The maps give the likelihood of a voxel takes a specific response magnitude for a contrast of interest with no reference to any specific information from other contrasts.