diff --git a/autoload/vimtex/fold.vim b/autoload/vimtex/fold.vim
index 10ff06a..1ef3dae 100644
--- a/autoload/vimtex/fold.vim
+++ b/autoload/vimtex/fold.vim
@@ -229,10 +229,8 @@ function! vimtex#fold#text() " {{{1
     let title =  matchstr(line, secpat2 . '\zs.*\ze\]')
   elseif line =~ secpat2
     let title =  matchstr(line, secpat2 . '\zs.*')
-  elseif line =~ 'Fake' . sections . ':'
-    let title =  matchstr(line,'Fake' . sections . ':\s*\zs.*')
   elseif line =~ 'Fake' . sections
-    let title =  matchstr(line, 'Fake' . sections)
+    let title =  matchstr(line, 'Fake' . sections . '.*')
   endif
 
   " Environments
diff --git a/doc/vimtex.txt b/doc/vimtex.txt
index 889f5a1..5bd0fc5 100644
--- a/doc/vimtex.txt
+++ b/doc/vimtex.txt
@@ -1360,6 +1360,18 @@ frontmatter, backmatter, and appendix) and section types (chapter, section,
 etc.) are present. This parsing is done automatically every time the folds are
 recomputed, if there are any changes to the file.
 
+The fold function also recognizes "fake" sections. That is, it parses comments
+similar to: >
+
+  % Fakepart title
+  % Fakechapter title
+  % Fakesection title
+  % Fakesubsection title
+
+The fake sections are folded at the same level as the corresponding "real"
+sections. The fold title is the provided title with the `Fake...` part
+prepended.
+
 Note: The |fold-expr| method of folding is well known to be slow, e.g. for
       long lines and large files. To speed things up, the user may want to
       enable the |g:vimtex_fold_manual| option. An alternative is to add