diff --git a/mne/parallel.py b/mne/parallel.py
index 225f49e313a..95798a52e7f 100644
--- a/mne/parallel.py
+++ b/mne/parallel.py
@@ -51,14 +51,33 @@ def parallel_func(func, n_jobs, verbose=None):
     parallel_verbose = 5 if logger.level <= logging.INFO else 0
     parallel = Parallel(n_jobs, verbose=parallel_verbose)
     my_func = delayed(func)
+    n_jobs = check_n_jobs(n_jobs)
+    return parallel, my_func, n_jobs
 
-    if n_jobs == -1:
-        try:
-            import multiprocessing
-            n_jobs = multiprocessing.cpu_count()
-        except ImportError:
-            logger.warn('multiprocessing not installed. Cannot run in '
-                         'parallel.')
-            n_jobs = 1
 
-    return parallel, my_func, n_jobs
+def check_n_jobs(n_jobs):
+    """Check n_jobs in particular for negative values
+
+    Parameters
+    ----------
+    n_jobs : int
+        The number of jobs
+
+    Returns
+    -------
+    n_jobs : int
+        The checked number of jobs. Always positive.
+    """
+    try:
+        import multiprocessing
+        n_cores = multiprocessing.cpu_count()
+        if n_cores + n_jobs <= 0:
+            raise ValueError('If n_jobs has a negative value it must not be less '
+                             'than the number of CPUs present. You\'ve got '
+                             '%s CPUs' % n_cores)
+        n_jobs = n_cores + n_jobs
+    except ImportError:
+        logger.warn('multiprocessing not installed. Cannot run in '
+                     'parallel.')
+        n_jobs = 1
+    return n_jobs
diff --git a/mne/stats/cluster_level.py b/mne/stats/cluster_level.py
index 8ad485dbc85..58bd0c7a5d5 100755
--- a/mne/stats/cluster_level.py
+++ b/mne/stats/cluster_level.py
@@ -9,14 +9,13 @@
 # License: Simplified BSD
 
 import numpy as np
-from multiprocessing import cpu_count
 from scipy import stats, sparse, ndimage
 
 import logging
 logger = logging.getLogger('mne')
 
 from .parametric import f_oneway
-from ..parallel import parallel_func
+from ..parallel import parallel_func, check_n_jobs
 from ..utils import split_list
 from ..fixes import in1d, unravel_index
 from .. import verbose
@@ -768,16 +767,7 @@ def permutation_cluster_1samp_test(X, threshold=1.67, n_permutations=1024,
     Journal of Neuroscience Methods, Vol. 164, No. 1., pp. 177-190.
     doi:10.1016/j.jneumeth.2007.03.024
     """
-
-    # infer number of jobs and replace if user passes negative integer
-    if n_jobs <= 0:
-        n_cores = cpu_count()
-        if n_cores + n_jobs <= 0:
-            raise ValueError('If n_jobs has a negative value it must not be less '
-                             'than the number of CPUs present. You\'ve got '
-                             '%s CPUs' % n_cores)
-        else:
-            n_jobs = n_cores + n_jobs
+    n_jobs = check_n_jobs(n_jobs)
 
     if not out_type in ['mask', 'indices']:
         raise ValueError('out_type must be either \'mask\' or \'indices\'')
diff --git a/mne/stats/tests/test_cluster_level.py b/mne/stats/tests/test_cluster_level.py
index 7055f2b9cf3..7503d132d67 100644
--- a/mne/stats/tests/test_cluster_level.py
+++ b/mne/stats/tests/test_cluster_level.py
@@ -8,8 +8,6 @@
                                     permutation_cluster_1samp_test, \
                                     spatio_temporal_cluster_1samp_test
 
-from multiprocessing import cpu_count
-
 noiselevel = 20
 
 normfactor = np.hanning(20).sum()
@@ -170,21 +168,6 @@ def test_cluster_permutation_t_test_with_connectivity():
                              connectivity=connectivity, max_step=1,
                              threshold=1.67, n_jobs=-1000)
 
-    # test negative n_jobs arg
-    n_jobs = -(cpu_count() - 1)
-    n_jobs2 = 1
-    out_1 = spatio_temporal_cluster_1samp_test(
-                             condition1_3, n_permutations=1,
-                             connectivity=connectivity, max_step=1,
-                             n_jobs=n_jobs,
-                             threshold=1.67)
-    out_2 = spatio_temporal_cluster_1samp_test(
-                             condition1_3, n_permutations=1,
-                             connectivity=connectivity, max_step=1,
-                             n_jobs=n_jobs2,
-                             threshold=1.67)
-    assert_array_almost_equal(out_1[2], out_2[2])
-
 
 def ttest_1samp(X):
     """Returns T-values