SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2501.17207 · 2025

Rethinking Functional Brain Connectome Analysis: Do Graph Deep Learning Models Help

Carl Yang, Vince Calhoun, Sergey Plis, Lifang He, Liang Zhan, Yao Su, Keqi Han

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2501.17207.json · badge: /api/ran/2501.17207.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
DCAN-Labs/abcd-hcp-pipeline canonical (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
FunctionStatusWhere it lives
set_anatomicals Ran
ran on a synthesized input
DCAN-Labs/abcd-hcp-pipeline/app/helpers.py
code served (permissive licence) · get_code("20ecb29b67388599")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

class MockBIDSFile:
    def __init__(self, path):
        self.path = path

class MockBIDSLayout:
    def get(self, subject=None, session=None, datatype=None, suffix=None, extension=None):
        if suffix == 'T1w':
            return [
                MockBIDSFile('/data/sub-01/ses-01/anat/sub-01_ses-01_T1w.nii.gz'),
                MockBIDSFile('/data/sub-01/ses-02/anat/sub-01_ses-02_T1w.nii.gz'),
            ]
        elif suffix == 'T2w':
            return [
                MockBIDSFile('/data/sub-01/ses-01/anat/sub-01_ses-01_T2w.nii.gz'),
            ]
        return []
    
    def get_metadata(self, path):
        return {
            'RepetitionTime': 2.3,
            'EchoTime': 0.00456,
            'FlipAngle': 9,
        }

layout = MockBIDSLayout()
subject = '01'
sessions = ['01', '02']

ARGS = {
    'layout': layout,
    'subject': subject,
    'sessions': sessions,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
set_fieldmaps Ran
ran on a synthesized input
DCAN-Labs/abcd-hcp-pipeline/app/helpers.py
code served (permissive licence) · get_code("5ea9303619859f12")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

class MockBIDSFile:
    def __init__(self, subject, session, suffix, path, metadata):
        self.path = path
        self.entities = {
            'subject': subject,
            'session': session,
            'suffix': suffix
        }
        self._metadata = metadata
    
    def get_metadata(self):
        return self._metadata


class MockBIDSLayout:
    def __init__(self, files):
        self.files = files
    
    def get(self, subject=None, session=None, datatype=None, suffix=None, extension=None):
        result = []
        for f in self.files:
            if subject and f.entities['subject'] != subject:
                continue
            if session and f.entities['session'] not in session:
                continue
            if suffix and f.entities['suffix'] not in suffix:
                continue
            result.append(f)
        return result


subject = "01"
sessions = ["01", "02"]

file1 = MockBIDSFile(
    subject="01",
    session="01",
    suffix="epi",
    path="/data/sub-01/ses-01/fmap/sub-01_ses-01_epi.nii.gz",
    metadata={
        'IntendedFor': ['sub-01_task-rest_bold.nii.gz'],
        'PhaseEncodingDirection': 'j'
    }
)

file2 = MockBIDSFile(
    subject="01",
    session="01",
    suffix="epi",
    path="/data/sub-01/ses-01/fmap/sub-01_ses-01_epi_rev.nii.gz",
    metadata={
        'IntendedFor': ['sub-01_task-rest_bold.nii.gz'],
        'PhaseEncodingDirection': 'j-'
    }
)

layout = MockBIDSLayout([file1, file2])

ARGS = {
    'layout': layout,
    'subject': subject,
    'sessions': sessions
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
set_functionals Ran
ran on a synthesized input
DCAN-Labs/abcd-hcp-pipeline/app/helpers.py
code served (permissive licence) · get_code("7e5ea0ab6d3af297")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np
from unittest.mock import Mock, MagicMock

# Mock BIDSLayout class with necessary methods
class MockBIDSFile:
    def __init__(self, path, suffix='bold'):
        self.path = path
        self.entities = {'suffix': suffix}

class MockBIDSLayout:
    def __init__(self):
        self.files = [
            MockBIDSFile('/data/sub-01/ses-01/func/sub-01_ses-01_bold.nii.gz', suffix='bold'),
            MockBIDSFile('/data/sub-01/ses-02/func/sub-01_ses-02_bold.nii.gz', suffix='bold'),
        ]
    
    def get(self, subject=None, session=None, datatype=None, suffix=None, extension=None):
        """Mock get method that filters and returns BIDSFile objects"""
        result = []
        for f in self.files:
            # Basic filtering logic
            if subject and 'sub-01' not in f.path:
                continue
            if session and not any(f'ses-{s}' in f.path for s in (session if isinstance(session, list) else [session])):
                continue
            if datatype and datatype not in f.path:
                continue
            if suffix and f.entities['suffix'] != suffix:
                continue
            result.append(f)
        return result
    
    def get_metadata(self, path):
        """Mock get_metadata method that returns metadata dict"""
        return {
            'RepetitionTime': 2.0,
            'EchoTime': 0.03,
            'FlipAngle': 90,
            'TaskName': 'rest'
        }

# Create layout instance
layout = MockBIDSLayout()

# Define arguments
ARGS = {
    'layout': layout,
    'subject': '01',
    'sessions': ['01', '02']
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 3 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Graph deep learning models, a class of AI-driven approaches employing a message aggregation mechanism, have gained popularity for analyzing the functional brain connectome in neuroimaging. However, their actual effectiveness remains unclear. In this study, we re-examine graph deep learning versus classical machine learning models based on four large-scale neuroimaging studies. Surprisingly, we find that the message aggregation mechanism, a hallmark of graph deep learning models, does not help with predictive performance as typically assumed, but rather consistently degrades it. To address this issue, we propose a hybrid model combining a linear model with a graph attention network through dual pathways, achieving robust predictions and enhanced interpretability by revealing both localized and global neural connectivity patterns. Our findings urge caution in adopting complex deep learning models for functional brain connectome analysis, emphasizing the need for rigorous experimental designs to establish tangible performance gains and perhaps more importantly, to pursue improvements in model interpretability.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2501.17207")
get_code_for_paper("2501.17207")
have("2501.17207")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2501.17207.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2501.17207.svg)](https://syntology.ai/paper/2501.17207)

Connect an agent — have() is free.