diff --git a/test/unittests/utils/test_atlas.py b/test/unittests/utils/test_atlas.py
index aa4288c1f..ccb74538d 100644
--- a/test/unittests/utils/test_atlas.py
+++ b/test/unittests/utils/test_atlas.py
@@ -1,4 +1,6 @@
 import re
+from pathlib import Path
+from typing import List
 
 import nibabel as nib
 import numpy as np
@@ -110,6 +112,7 @@ def test_atlas_factory(tmp_path, monkeypatch, atlas_name, atlas):
     ),
 )
 def test_atlases(
+    tmp_path,
     atlas,
     expected_name,
     expected_checksum,
@@ -117,7 +120,16 @@ def test_atlases(
     expected_roi_filename,
     expected_resolution,
     expected_size,
+    mocker,
 ):
+    # The following atlases are supposed to be downloaded from a remote server
+    # For unit testing, we mock the get_file_from_server function to return
+    # a nifti image with fake data realistic enough to make the tests pass
+    if expected_name in ("Hammers", "LPBA40", "Neuromorphometrics"):
+        mocker.patch(
+            "clinica.utils.inputs.get_file_from_server",
+            return_value=get_mocked_atlas(tmp_path, atlas),
+        )
     assert atlas.name == expected_name
     assert atlas.expected_checksum == expected_checksum
     assert atlas.atlas_filename == expected_atlas_filename
@@ -128,6 +140,107 @@ def test_atlases(
     assert_array_equal(atlas.get_index(), np.arange(expected_size))
 
 
+def get_mocked_atlas(tmp_path, atlas) -> Path:
+    """Return the path to the mocked atlas label image.
+
+    We need the mocked image to have:
+        - the right resolution (same for all mocked atlases)
+        - the right data shape (same for all mocked atlases)
+        - the right label values
+
+    To ensure the last point, this function generate a data array
+    composed of the cycled expected labels.
+    """
+    from itertools import cycle
+
+    data_shape = (121, 145, 121)
+    gen = cycle(get_labels(atlas))
+    mocked_data = []
+    while len(mocked_data) < np.prod(data_shape):
+        mocked_data.append(next(gen))
+    mocked_data = np.array(mocked_data, dtype="float")
+    mocked_data = mocked_data.reshape(data_shape)
+    mocked_image = nib.Nifti1Image(mocked_data, np.diag([-1.5, 1.5, 1.5, 1.0]))
+    nib.save(mocked_image, tmp_path / "mocked_atlas.nii.gz")
+
+    return tmp_path / "mocked_atlas.nii.gz"
+
+
+def get_labels(atlas) -> List[int]:
+    """Return the expected label values for the mocked atlases."""
+    if atlas.name == "Neuromorphometrics":
+        labels = list(range(143))
+        labels.remove(35)
+        labels.remove(23)
+    elif atlas.name == "LPBA40":
+        labels = [
+            0,
+            21,
+            22,
+            23,
+            24,
+            25,
+            26,
+            27,
+            28,
+            29,
+            30,
+            31,
+            32,
+            33,
+            34,
+            41,
+            42,
+            43,
+            44,
+            45,
+            46,
+            47,
+            48,
+            49,
+            50,
+            61,
+            62,
+            63,
+            64,
+            65,
+            66,
+            67,
+            68,
+            81,
+            82,
+            83,
+            84,
+            85,
+            86,
+            87,
+            88,
+            89,
+            90,
+            91,
+            92,
+            101,
+            102,
+            121,
+            122,
+            161,
+            162,
+            163,
+            164,
+            165,
+            166,
+            181,
+            182,
+        ]
+    elif atlas.name == "Hammers":
+        labels = list(range(69))
+    else:
+        raise ValueError(
+            f"Atlas {atlas.name} is not supposed to be mocked in this test."
+        )
+    return labels
+
+
 @pytest.fixture
 def atlas(expected_name, tmp_path, monkeypatch):
     from clinica.utils.atlas import atlas_factory