diff --git a/nibabel/filebasedimages.py b/nibabel/filebasedimages.py
index 86ce837942..90bbd8e652 100644
--- a/nibabel/filebasedimages.py
+++ b/nibabel/filebasedimages.py
@@ -246,7 +246,7 @@ def set_filename(self, filename):
 
         Parameters
         ----------
-        filename : str
+        filename : str or os.PathLike
            If the image format only has one file associated with it,
            this will be the only filename set into the image
            ``.file_map`` attribute. Otherwise, the image instance will
@@ -279,7 +279,7 @@ def filespec_to_file_map(klass, filespec):
 
         Parameters
         ----------
-        filespec : str
+        filespec : str or os.PathLike
             Filename that might be for this image file type.
 
         Returns
@@ -321,7 +321,7 @@ def to_filename(self, filename):
 
         Parameters
         ----------
-        filename : str
+        filename : str or os.PathLike
            filename to which to save image.  We will parse `filename`
            with ``filespec_to_file_map`` to work out names for image,
            header etc.
@@ -419,7 +419,7 @@ def _sniff_meta_for(klass, filename, sniff_nbytes, sniff=None):
 
         Parameters
         ----------
-        filename : str
+        filename : str or os.PathLike
             Filename for an image, or an image header (metadata) file.
             If `filename` points to an image data file, and the image type has
             a separate "header" file, we work out the name of the header file,
@@ -466,7 +466,7 @@ def path_maybe_image(klass, filename, sniff=None, sniff_max=1024):
 
         Parameters
         ----------
-        filename : str
+        filename : str or os.PathLike
             Filename for an image, or an image header (metadata) file.
             If `filename` points to an image data file, and the image type has
             a separate "header" file, we work out the name of the header file,
diff --git a/nibabel/filename_parser.py b/nibabel/filename_parser.py
index db6e073018..ed04610fdd 100644
--- a/nibabel/filename_parser.py
+++ b/nibabel/filename_parser.py
@@ -9,16 +9,43 @@
 ''' Create filename pairs, triplets etc, with expected extensions '''
 
 import os
-try:
-    basestring
-except NameError:
-    basestring = str
+import pathlib
 
 
 class TypesFilenamesError(Exception):
     pass
 
 
+def _stringify_path(filepath_or_buffer):
+    """Attempt to convert a path-like object to a string.
+
+    Parameters
+    ----------
+    filepath_or_buffer : str or os.PathLike
+
+    Returns
+    -------
+    str_filepath_or_buffer : str
+
+    Notes
+    -----
+    Objects supporting the fspath protocol (python 3.6+) are coerced
+    according to its __fspath__ method.
+    For backwards compatibility with older pythons, pathlib.Path objects
+    are specially coerced.
+    Any other object is passed through unchanged, which includes bytes,
+    strings, buffers, or anything else that's not even path-like.
+
+    Copied from:
+    https://github.com/pandas-dev/pandas/blob/325dd686de1589c17731cf93b649ed5ccb5a99b4/pandas/io/common.py#L131-L160
+    """
+    if hasattr(filepath_or_buffer, "__fspath__"):
+        return filepath_or_buffer.__fspath__()
+    elif isinstance(filepath_or_buffer, pathlib.Path):
+        return str(filepath_or_buffer)
+    return filepath_or_buffer
+
+
 def types_filenames(template_fname, types_exts,
                     trailing_suffixes=('.gz', '.bz2'),
                     enforce_extensions=True,
@@ -31,7 +58,7 @@ def types_filenames(template_fname, types_exts,
 
     Parameters
     ----------
-    template_fname : str
+    template_fname : str or os.PathLike
        template filename from which to construct output dict of
        filenames, with given `types_exts` type to extension mapping.  If
        ``self.enforce_extensions`` is True, then filename must have one
@@ -82,7 +109,8 @@ def types_filenames(template_fname, types_exts,
     >>> tfns == {'t1': '/path/test.funny', 't2': '/path/test.ext2'}
     True
     '''
-    if not isinstance(template_fname, basestring):
+    template_fname = _stringify_path(template_fname)
+    if not isinstance(template_fname, str):
         raise TypesFilenamesError('Need file name as input '
                                   'to set_filenames')
     if template_fname.endswith('.'):
@@ -151,7 +179,7 @@ def parse_filename(filename,
 
     Parameters
     ----------
-    filename : str
+    filename : str or os.PathLike
        filename in which to search for type extensions
     types_exts : sequence of sequences
        sequence of (name, extension) str sequences defining type to
@@ -190,6 +218,8 @@ def parse_filename(filename,
     >>> parse_filename('/path/fnameext2.gz', types_exts, ('.gz',))
     ('/path/fname', 'ext2', '.gz', 't2')
     '''
+    filename = _stringify_path(filename)
+
     ignored = None
     if match_case:
         endswith = _endswith
@@ -232,7 +262,7 @@ def splitext_addext(filename,
 
     Parameters
     ----------
-    filename : str
+    filename : str or os.PathLike
        filename that may end in any or none of `addexts`
     match_case : bool, optional
        If True, match case of `addexts` and `filename`, otherwise do
@@ -257,6 +287,8 @@ def splitext_addext(filename,
     >>> splitext_addext('fname.ext.foo', ('.foo', '.bar'))
     ('fname', '.ext', '.foo')
     '''
+    filename = _stringify_path(filename)
+
     if match_case:
         endswith = _endswith
     else:
diff --git a/nibabel/freesurfer/mghformat.py b/nibabel/freesurfer/mghformat.py
index ddb30cb796..6eb0f156e9 100644
--- a/nibabel/freesurfer/mghformat.py
+++ b/nibabel/freesurfer/mghformat.py
@@ -17,6 +17,7 @@
 from ..volumeutils import (array_to_file, array_from_file, endian_codes,
                            Recoder)
 from ..filebasedimages import SerializableImage
+from ..filename_parser import _stringify_path
 from ..spatialimages import HeaderDataError, SpatialImage
 from ..fileholders import FileHolder
 from ..arrayproxy import ArrayProxy, reshape_dataobj
@@ -529,6 +530,7 @@ def __init__(self, dataobj, affine, header=None,
 
     @classmethod
     def filespec_to_file_map(klass, filespec):
+        filespec = _stringify_path(filespec)
         """ Check for compressed .mgz format, then .mgh format """
         if splitext(filespec)[1].lower() == '.mgz':
             return dict(image=FileHolder(filename=filespec))
diff --git a/nibabel/loadsave.py b/nibabel/loadsave.py
index d603dd619c..f8c3e3be0b 100644
--- a/nibabel/loadsave.py
+++ b/nibabel/loadsave.py
@@ -12,7 +12,7 @@
 import os
 import numpy as np
 
-from .filename_parser import splitext_addext
+from .filename_parser import splitext_addext, _stringify_path
 from .openers import ImageOpener
 from .filebasedimages import ImageFileError
 from .imageclasses import all_image_classes
@@ -25,7 +25,7 @@ def load(filename, **kwargs):
 
     Parameters
     ----------
-    filename : string
+    filename : str or os.PathLike
        specification of file to load
     \*\*kwargs : keyword arguments
         Keyword arguments to format-specific load
@@ -35,12 +35,16 @@ def load(filename, **kwargs):
     img : ``SpatialImage``
        Image of guessed type
     '''
+    filename = _stringify_path(filename)
+
+    # Check file exists and is not empty
     try:
         stat_result = os.stat(filename)
     except OSError:
         raise FileNotFoundError("No such file or no access: '%s'" % filename)
     if stat_result.st_size <= 0:
         raise ImageFileError("Empty file: '%s'" % filename)
+
     sniff = None
     for image_klass in all_image_classes:
         is_valid, sniff = image_klass.path_maybe_image(filename, sniff)
@@ -85,13 +89,14 @@ def save(img, filename):
     ----------
     img : ``SpatialImage``
        image to save
-    filename : str
+    filename : str or os.PathLike
        filename (often implying filenames) to which to save `img`.
 
     Returns
     -------
     None
     '''
+    filename = _stringify_path(filename)
 
     # Save the type as expected
     try:
diff --git a/nibabel/tests/test_image_api.py b/nibabel/tests/test_image_api.py
index 748f9c2472..3b921b9fb9 100644
--- a/nibabel/tests/test_image_api.py
+++ b/nibabel/tests/test_image_api.py
@@ -26,6 +26,7 @@
 import warnings
 from functools import partial
 from itertools import product
+import pathlib
 
 import numpy as np
 
@@ -141,21 +142,23 @@ def validate_filenames(self, imaker, params):
         assert_almost_equal(np.asanyarray(img.dataobj), np.asanyarray(rt_img.dataobj))
         # get_ / set_ filename
         fname = 'an_image' + self.standard_extension
-        img.set_filename(fname)
-        assert_equal(img.get_filename(), fname)
-        assert_equal(img.file_map['image'].filename, fname)
+        for path in (fname, pathlib.Path(fname)):
+            img.set_filename(path)
+            assert_equal(img.get_filename(), str(path))
+            assert_equal(img.file_map['image'].filename, str(path))
         # to_ / from_ filename
         fname = 'another_image' + self.standard_extension
-        with InTemporaryDirectory():
-            # Validate that saving or loading a file doesn't use deprecated methods internally
-            with clear_and_catch_warnings() as w:
-                warnings.simplefilter('error', DeprecationWarning)
-                img.to_filename(fname)
-                rt_img = img.__class__.from_filename(fname)
-            assert_array_equal(img.shape, rt_img.shape)
-            assert_almost_equal(img.get_fdata(), rt_img.get_fdata())
-            assert_almost_equal(np.asanyarray(img.dataobj), np.asanyarray(rt_img.dataobj))
-            del rt_img  # to allow windows to delete the directory
+        for path in (fname, pathlib.Path(fname)):
+          with InTemporaryDirectory():
+              # Validate that saving or loading a file doesn't use deprecated methods internally
+              with clear_and_catch_warnings() as w:
+                  warnings.simplefilter('error', DeprecationWarning)
+                  img.to_filename(path)
+                  rt_img = img.__class__.from_filename(path)
+              assert_array_equal(img.shape, rt_img.shape)
+              assert_almost_equal(img.get_fdata(), rt_img.get_fdata())
+              assert_almost_equal(np.asanyarray(img.dataobj), np.asanyarray(rt_img.dataobj))
+              del rt_img  # to allow windows to delete the directory
 
     def validate_no_slicing(self, imaker, params):
         img = imaker()
diff --git a/nibabel/tests/test_image_load_save.py b/nibabel/tests/test_image_load_save.py
index 6031d4e851..9d58a3ed60 100644
--- a/nibabel/tests/test_image_load_save.py
+++ b/nibabel/tests/test_image_load_save.py
@@ -12,6 +12,7 @@
 import shutil
 from os.path import dirname, join as pjoin
 from tempfile import mkdtemp
+import pathlib
 
 import numpy as np
 
@@ -255,13 +256,14 @@ def test_filename_save():
         try:
             pth = mkdtemp()
             fname = pjoin(pth, 'image' + out_ext)
-            nils.save(img, fname)
-            rt_img = nils.load(fname)
-            assert_array_almost_equal(rt_img.get_fdata(), data)
-            assert_true(type(rt_img) is loadklass)
-            # delete image to allow file close.  Otherwise windows
-            # raises an error when trying to delete the directory
-            del rt_img
+            for path in (fname, pathlib.Path(fname)):
+                nils.save(img, path)
+                rt_img = nils.load(path)
+                assert_array_almost_equal(rt_img.get_fdata(), data)
+                assert_true(type(rt_img) is loadklass)
+                # delete image to allow file close.  Otherwise windows
+                # raises an error when trying to delete the directory
+                del rt_img
         finally:
             shutil.rmtree(pth)
 
diff --git a/nibabel/tests/test_loadsave.py b/nibabel/tests/test_loadsave.py
index 491cb07b76..3d7101b6d3 100644
--- a/nibabel/tests/test_loadsave.py
+++ b/nibabel/tests/test_loadsave.py
@@ -3,6 +3,7 @@
 
 from os.path import dirname, join as pjoin
 import shutil
+import pathlib
 
 import numpy as np
 
@@ -26,14 +27,20 @@
 
 
 def test_read_img_data():
-    for fname in ('example4d.nii.gz',
-                  'example_nifti2.nii.gz',
-                  'minc1_1_scale.mnc',
-                  'minc1_4d.mnc',
-                  'test.mgz',
-                  'tiny.mnc'
-                  ):
-        fpath = pjoin(data_path, fname)
+    fnames_test = [
+        'example4d.nii.gz',
+        'example_nifti2.nii.gz',
+        'minc1_1_scale.mnc',
+        'minc1_4d.mnc',
+        'test.mgz',
+        'tiny.mnc'
+    ]
+    fnames_test += [pathlib.Path(p) for p in fnames_test]
+    for fname in fnames_test:
+        # os.path.join doesnt work between str / os.PathLike in py3.5
+        fpath = pjoin(data_path, str(fname))
+        if isinstance(fname, pathlib.Path):
+            fpath = pathlib.Path(fpath)
         img = load(fpath)
         data = img.get_fdata()
         data2 = read_img_data(img)
@@ -45,8 +52,11 @@ def test_read_img_data():
             assert_array_equal(read_img_data(img, prefer='unscaled'), data)
         # Assert all caps filename works as well
         with TemporaryDirectory() as tmpdir:
-            up_fpath = pjoin(tmpdir, fname.upper())
-            shutil.copyfile(fpath, up_fpath)
+            up_fpath = pjoin(tmpdir, str(fname).upper())
+            if isinstance(fname, pathlib.Path):
+                up_fpath = pathlib.Path(up_fpath)
+            # shutil doesnt work with os.PathLike in py3.5
+            shutil.copyfile(str(fpath), str(up_fpath))
             img = load(up_fpath)
             assert_array_equal(img.dataobj, data)
             del img