diff --git a/CMakeLists.txt b/CMakeLists.txt
index 4d7122ef..a5aa1fc3 100755
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -152,7 +152,7 @@ if(USE_OPENCL)
     message(STATUS "Found OpenCL")
     include_directories(${CMAKE_SOURCE_DIR}/reg-lib/cl)
     include_directories(${OpenCL_INCLUDE_DIRS})
-    add_definitions(-D_USE_OPENCL)
+    add_definitions(-DUSE_OPENCL)
   endif(NOT OpenCL_FOUND)
 endif(USE_OPENCL)
 #-----------------------------------------------------------------------------
@@ -168,7 +168,7 @@ if(USE_CUDA)
   else(NOT CUDA_FOUND)
     include_directories(${CMAKE_SOURCE_DIR}/reg-lib/cuda)
     include_directories(${CUDA_INCLUDE_DIRS})
-    add_definitions(-D_USE_CUDA)
+    add_definitions(-DUSE_CUDA)
   endif(NOT CUDA_FOUND)
 endif(USE_CUDA)
 #-----------------------------------------------------------------------------
@@ -177,7 +177,7 @@ if(USE_SSE)
     set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -msse3")
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msse3")
   endif(NOT MSVC)
-  add_definitions(-D_USE_SSE)
+  add_definitions(-DUSE_SSE)
 endif(USE_SSE)
 #-----------------------------------------------------------------------------
 if(USE_OPENMP)
diff --git a/Doxyfile.in b/Doxyfile.in
index 8257d784..df013886 100644
--- a/Doxyfile.in
+++ b/Doxyfile.in
@@ -1449,7 +1449,7 @@ INCLUDE_FILE_PATTERNS  =
 # undefined via #undef or recursively expanded use the := operator
 # instead of the = operator.
 
-PREDEFINED             = _USE_CUDA
+PREDEFINED             = USE_CUDA
 
 # If the MACRO_EXPANSION and EXPAND_ONLY_PREDEF tags are set to YES then
 # this tag can be used to specify a list of macro names that should be expanded.
diff --git a/niftyreg_build_version.txt b/niftyreg_build_version.txt
index 71627d71..aef2e272 100644
--- a/niftyreg_build_version.txt
+++ b/niftyreg_build_version.txt
@@ -1 +1 @@
-348
+349
diff --git a/reg-apps/reg_benchmark.cpp b/reg-apps/reg_benchmark.cpp
index 8606f563..aab0086c 100644
--- a/reg-apps/reg_benchmark.cpp
+++ b/reg-apps/reg_benchmark.cpp
@@ -19,7 +19,7 @@
 #include "_reg_tools.h"
 #include "_reg_blockMatching.h"
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
 #include "_reg_cudaCommon.h"
 #include "_reg_resampling_gpu.h"
 #include "_reg_affineTransformation_gpu.h"
@@ -179,7 +179,7 @@ int main(int argc, char **argv)
    nodeNMIGradientImage->nbyper = sizeof(float);
    nodeNMIGradientImage->data = calloc(nodeNMIGradientImage->nvox, nodeNMIGradientImage->nbyper);
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    float *targetImageArray_d;
    cudaArray *sourceImageArray_d;
    int *targetMask_d;
@@ -198,7 +198,7 @@ int main(int argc, char **argv)
 
    time_t start,end;
    int minutes, seconds, cpuTime, maxIt;
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    int gpuTime
 #endif
 
@@ -249,7 +249,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf( "CPU - %i affine deformation field computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i affine deformation field computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -273,7 +273,7 @@ int main(int argc, char **argv)
    }
 
    // SPLINE DEFORMATION FIELD CREATION
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    float4 *controlPointImageArray_d;
    if(runGPU)
    {
@@ -299,7 +299,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i spline deformation field computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i spline deformation field computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -326,7 +326,7 @@ int main(int argc, char **argv)
    }
 
    // SCALING-AND-SQUARING APPROACH
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    float4 *velocityFieldImageArray_d;
    if(runGPU)
    {
@@ -350,7 +350,7 @@ int main(int argc, char **argv)
       printf("CPU - %i scaling-and-squaring - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i scaling-and-squarings - %i min %i sec\n", maxIt, minutes, seconds);
       time(&start);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          for(int i=0; i<maxIt; ++i)
@@ -374,7 +374,7 @@ int main(int argc, char **argv)
    }
 
    // LINEAR INTERPOLATION
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    float *resultImageArray_d;
    if(runGPU)
       Cuda::Allocate<float>(&resultImageArray_d, targetImage->dim);
@@ -399,7 +399,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i linear interpolation computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i linear interpolation computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -428,7 +428,7 @@ int main(int argc, char **argv)
    }
 
    // SPATIAL GRADIENT COMPUTATION
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    float4 *resultGradientArray_d;
    CUDA_SAFE_CALL(cudaMalloc((void **)&resultGradientArray_d, targetImage->nvox*sizeof(float4)));
 #endif
@@ -451,7 +451,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i spatial gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i spatial gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -479,7 +479,7 @@ int main(int argc, char **argv)
    }
    nifti_image_free(sourceImage);
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    if(runGPU)
    {
       Cuda::Free(deformationFieldImageArray_d);
@@ -501,7 +501,7 @@ int main(int argc, char **argv)
    }
 
    // VOXEL-BASED NMI GRADIENT COMPUTATION
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    float4 *voxelNMIGradientArray_d;
    if(runGPU)
       Cuda::Allocate(&voxelNMIGradientArray_d, resultImage->dim);
@@ -529,7 +529,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i voxel-based NMI gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i voxel-based NMI gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       float *logJointHistogram_d;
       if(runGPU)
       {
@@ -571,7 +571,7 @@ int main(int argc, char **argv)
       printf("Voxel-based NMI gradient done\n\n");
    }
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    if(runGPU)
    {
       Cuda::Free(resultGradientArray_d);
@@ -579,7 +579,7 @@ int main(int argc, char **argv)
 #endif
 
    // NODE-BASED NMI GRADIENT COMPUTATION
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    float4 *nodeNMIGradientArray_d;
    if(runGPU)
       Cuda::Allocate(&nodeNMIGradientArray_d, controlPointImage->dim);
@@ -603,7 +603,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i node-based NMI gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i node-based NMI gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -631,7 +631,7 @@ int main(int argc, char **argv)
       printf("Node-based NMI gradient done\n\n");
    }
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    if(runGPU)
    {
       Cuda::Free(voxelNMIGradientArray_d);
@@ -654,7 +654,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i BE computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i BE computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -694,7 +694,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i BE gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i BE gradient computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -733,7 +733,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i |Jac| penalty term computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i |Jac| penalty term computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -768,7 +768,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i Approx. |Jac| penalty term computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i Approx. |Jac| penalty term computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -789,7 +789,7 @@ int main(int argc, char **argv)
       printf("Approx. |Jac| penalty term done\n\n");
    }
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    if(runGPU)
    {
       Cuda::Free(controlPointImageArray_d );
@@ -806,7 +806,7 @@ int main(int argc, char **argv)
                                           100,    // percentage of block kept
                                           50,     // percentage of inlier in the optimisation process
                                           maskImage);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       int *activeBlock_d;
       float *targetPosition_d;
       float *resultPosition_d;
@@ -835,7 +835,7 @@ int main(int argc, char **argv)
       seconds = (int)(cpuTime - 60*minutes);
       printf("CPU - %i block matching computations - %i min %i sec\n", maxIt, minutes, seconds);
       fprintf(outputFile, "CPU - %i block matching computations - %i min %i sec\n", maxIt, minutes, seconds);
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
       if(runGPU)
       {
          time(&start);
@@ -880,7 +880,7 @@ int main(int argc, char **argv)
    free(probaJointHistogram);
    free(logJointHistogram);
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    if(runGPU)
    {
       Cuda::Free(targetImageArray_d);
diff --git a/reg-apps/reg_f3d.cpp b/reg-apps/reg_f3d.cpp
index 104803a4..ad804dcd 100755
--- a/reg-apps/reg_f3d.cpp
+++ b/reg-apps/reg_f3d.cpp
@@ -11,7 +11,7 @@
  */
 
 // OpenCL isn't supported!
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "_reg_ReadWriteImage.h"
 #include "_reg_ReadWriteMatrix.h"
diff --git a/reg-apps/reg_gpuinfo.cpp b/reg-apps/reg_gpuinfo.cpp
index 22008d4e..d4858ead 100644
--- a/reg-apps/reg_gpuinfo.cpp
+++ b/reg-apps/reg_gpuinfo.cpp
@@ -1,25 +1,25 @@
 #include "_reg_maths.h"
 #include "Platform.h"
 
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
   #include "../reg-lib/cuda/_reg_cudainfo.h"
 #endif
-#ifdef _USE_OPENCL
+#ifdef USE_OPENCL
   #include "../reg-lib/cl/_reg_openclinfo.h"
 #endif
 
 /* *************************************************************** */
 int main()
 {
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
    showCUDAInfo();
 #else
-#ifndef _USE_OPENCL
+#ifndef USE_OPENCL
    NR_WARN("NiftyReg has not been compiled with CUDA or OpenCL");
    NR_WARN("No GPU device information to display");
 #endif
 #endif
-#ifdef _USE_OPENCL
+#ifdef USE_OPENCL
    showCLInfo();
 #endif
 
diff --git a/reg-io/CMakeLists.txt b/reg-io/CMakeLists.txt
index 82a541ca..639785ea 100644
--- a/reg-io/CMakeLists.txt
+++ b/reg-io/CMakeLists.txt
@@ -17,7 +17,7 @@ set(LIBRARIES reg_nifti reg_png)
 
 # Build the NRRD file format library if required
 if(USE_NRRD)
-  add_definitions(-D_USE_NRRD)
+  add_definitions(-DUSE_NRRD)
   subdirs(nrrd)
   set(LIBRARIES ${LIBRARIES} reg_nrrd)
 endif(USE_NRRD)
diff --git a/reg-io/_reg_ReadWriteImage.cpp b/reg-io/_reg_ReadWriteImage.cpp
index 4902881d..b5413b21 100644
--- a/reg-io/_reg_ReadWriteImage.cpp
+++ b/reg-io/_reg_ReadWriteImage.cpp
@@ -42,7 +42,7 @@ int reg_io_checkFileFormat(const std::string& filename) {
         return NR_NII_FORMAT;
     else if (filename.find(".png") != std::string::npos)
         return NR_PNG_FORMAT;
-#ifdef _USE_NRRD
+#ifdef USE_NRRD
     else if (filename.find(".nrrd") != std::string::npos)
         return NR_NRRD_FORMAT;
     else if (filename.find(".nhdr") != std::string::npos)
@@ -72,7 +72,7 @@ nifti_image* reg_io_ReadImageFile(const char *filename) {
         image = reg_io_readPNGfile(filename, true);
         reg_hack_filename(image, filename);
         break;
-#ifdef _USE_NRRD
+#ifdef USE_NRRD
     case NR_NRRD_FORMAT:
         Nrrd *nrrdImage = reg_io_readNRRDfile(filename);
         image = reg_io_nrdd2nifti(nrrdImage);
@@ -103,7 +103,7 @@ nifti_image* reg_io_ReadImageHeader(const char *filename) {
         image = reg_io_readPNGfile(filename, false);
         reg_hack_filename(image, filename);
         break;
-#ifdef _USE_NRRD
+#ifdef USE_NRRD
     case NR_NRRD_FORMAT:
         Nrrd *nrrdImage = reg_io_readNRRDfile(filename);
         image = reg_io_nrdd2nifti(nrrdImage);
@@ -154,7 +154,7 @@ void reg_io_WriteImageFile(nifti_image *image, const char *filename) {
     case NR_PNG_FORMAT:
         reg_io_writePNGfile(image, filename);
         break;
-#ifdef _USE_NRRD
+#ifdef USE_NRRD
     case NR_NRRD_FORMAT:
         Nrrd *nrrdImage = reg_io_nifti2nrrd(image);
         reg_io_writeNRRDfile(nrrdImage, filename);
diff --git a/reg-io/_reg_ReadWriteImage.h b/reg-io/_reg_ReadWriteImage.h
index a012f6c0..c1356f02 100644
--- a/reg-io/_reg_ReadWriteImage.h
+++ b/reg-io/_reg_ReadWriteImage.h
@@ -18,7 +18,7 @@
 #include "_reg_tools.h"
 
 #include "reg_png.h"
-#ifdef _USE_NRRD
+#ifdef USE_NRRD
 #include "reg_nrrd.h"
 #endif
 /** @defgroup NIFTYREG_FILEFORMAT_TYPE
@@ -27,7 +27,7 @@
  */
 #define NR_NII_FORMAT 0
 #define NR_PNG_FORMAT 1
-#ifdef _USE_NRRD
+#ifdef USE_NRRD
 #define NR_NRRD_FORMAT 2
 #endif
 /* @} */
diff --git a/reg-io/niftilib/nifti1_io.c b/reg-io/niftilib/nifti1_io.c
index afd444c9..d8bee4da 100644
--- a/reg-io/niftilib/nifti1_io.c
+++ b/reg-io/niftilib/nifti1_io.c
@@ -1,4 +1,4 @@
-#define _NIFTI1_IO_C_
+#define NIFTI1_IO_C
 
 #include "niftilib/nifti1_io.h"   /* typedefs, prototypes, macros, etc. */
 
@@ -192,7 +192,7 @@ static char const * const gni_history[] =
   "\n",
   "1.3  09 Feb 2005 [rickr]\n"
   "   - nifti1.h: added doxygen comments for extension structs\n"
-  "   - nifti1_io.h: put most #defines in #ifdef _NIFTI1_IO_C_ block\n"
+  "   - nifti1_io.h: put most #defines in #ifdef NIFTI1_IO_C block\n"
   "   - added a doxygen-style description to every exported function\n"
   "   - added doxygen-style comments within some functions\n"
   "   - re-exported many znzFile functions that I had made static\n"
@@ -7188,7 +7188,7 @@ static int make_pivot_list(nifti_image * nim, const int dims[], int pivots[],
    dim_index = nim->dim[0];
    while( dim_index > 0 ){
       prods[len] = 1;
-      while( dim_index > 0 && 
+      while( dim_index > 0 &&
              (nim->dim[dim_index] == 1 || dims[dim_index] == -1) ){
          prods[len] *= nim->dim[dim_index];
          dim_index--;
diff --git a/reg-io/niftilib/nifti1_io.h b/reg-io/niftilib/nifti1_io.h
index 14ed0d3a..0e95531c 100644
--- a/reg-io/niftilib/nifti1_io.h
+++ b/reg-io/niftilib/nifti1_io.h
@@ -517,9 +517,9 @@ int    valid_nifti_extensions(const nifti_image *nim);
 #endif
 
 /*------------------------------------------------------------------------*/
-/*-- the rest of these apply only to nifti1_io.c, check for _NIFTI1_IO_C_ */
+/*-- the rest of these apply only to nifti1_io.c, check for NIFTI1_IO_C */
 /*                                                    Feb 9, 2005 [rickr] */
-#ifdef _NIFTI1_IO_C_
+#ifdef NIFTI1_IO_C
 
 typedef struct {
     int debug;               /*!< debug level for status reports  */
@@ -574,7 +574,7 @@ typedef struct {
 
 #define LNI_MAX_NIA_EXT_LEN 100000  /* consider a longer extension invalid */
 
-#endif  /* _NIFTI1_IO_C_ section */
+#endif  /* NIFTI1_IO_C section */
 /*------------------------------------------------------------------------*/
 
 /*=================*/
diff --git a/reg-io/niftilib/nifti2_io.c b/reg-io/niftilib/nifti2_io.c
index da972895..a87fa3fd 100644
--- a/reg-io/niftilib/nifti2_io.c
+++ b/reg-io/niftilib/nifti2_io.c
@@ -1,4 +1,4 @@
-#define _NIFTI2_IO_C_
+#define NIFTI2_IO_C
 
 #include "niftilib/nifti2_io.h"   /* typedefs, prototypes, macros, etc. */
 
@@ -192,7 +192,7 @@ static char const * const gni1_history[] =
   "\n",
   "1.3  09 Feb 2005 [rickr]\n"
   "   - nifti1.h: added doxygen comments for extension structs\n"
-  "   - nifti1_io.h: put most #defines in #ifdef _NIFTI1_IO_C_ block\n"
+  "   - nifti1_io.h: put most #defines in #ifdef NIFTI1_IO_C block\n"
   "   - added a doxygen-style description to every exported function\n"
   "   - added doxygen-style comments within some functions\n"
   "   - re-exported many znzFile functions that I had made static\n"
diff --git a/reg-io/niftilib/nifti2_io.h b/reg-io/niftilib/nifti2_io.h
index ff215d19..946e6d4e 100644
--- a/reg-io/niftilib/nifti2_io.h
+++ b/reg-io/niftilib/nifti2_io.h
@@ -756,9 +756,9 @@ int    nifti_valid_header_size(int ni_ver, int whine);
 #endif
 
 /*------------------------------------------------------------------------*/
-/*-- the rest of these apply only to nifti2_io.c, check for _NIFTI2_IO_C_ */
+/*-- the rest of these apply only to nifti2_io.c, check for NIFTI2_IO_C */
 
-#ifdef _NIFTI2_IO_C_
+#ifdef NIFTI2_IO_C
 
 typedef struct {
     int debug;               /*!< debug level for status reports  */
@@ -817,7 +817,7 @@ typedef struct {
 #undef NIFTI_IS_16_BIT_INT
 #define NIFTI_IS_16_BIT_INT(x) ((x) <= 32767 && (x) >= -32768)
 
-#endif  /* _NIFTI2_IO_C_ section */
+#endif  /* NIFTI2_IO_C section */
 /*------------------------------------------------------------------------*/
 
 /*=================*/
diff --git a/reg-io/nrrd/NrrdIO/mangle.pl b/reg-io/nrrd/NrrdIO/mangle.pl
index 37c44fa9..f71c3299 100644
--- a/reg-io/nrrd/NrrdIO/mangle.pl
+++ b/reg-io/nrrd/NrrdIO/mangle.pl
@@ -2,23 +2,23 @@
 #  NrrdIO: stand-alone code for basic nrrd functionality
 #  Copyright (C) 2005  Gordon Kindlmann
 #  Copyright (C) 2004, 2003, 2002, 2001, 2000, 1999, 1998  University of Utah
-# 
+#
 #  This software is provided 'as-is', without any express or implied
 #  warranty.  In no event will the authors be held liable for any
 #  damages arising from the use of this software.
-# 
+#
 #  Permission is granted to anyone to use this software for any
 #  purpose, including commercial applications, and to alter it and
 #  redistribute it freely, subject to the following restrictions:
-# 
+#
 #  1. The origin of this software must not be misrepresented; you must
 #     not claim that you wrote the original software. If you use this
 #     software in a product, an acknowledgment in the product
 #     documentation would be appreciated but is not required.
-# 
+#
 #  2. Altered source versions must be plainly marked as such, and must
 #     not be misrepresented as being the original software.
-# 
+#
 #  3. This notice may not be removed or altered from any source distribution.
 #
 #
@@ -41,8 +41,7 @@
     $mac = 0;
 }
 
-print "#ifndef __${prefix}_NrrdIO_mangle_h\n";
-print "#define __${prefix}_NrrdIO_mangle_h\n";
+print "#pragma once\n";
 print "\n";
 print "/*\n";
 print "\n";
@@ -89,4 +88,3 @@
     }
 }
 close(NM);
-print "#endif  /* __${prefix}_NrrdIO_mangle_h */ \n";
diff --git a/reg-lib/Platform.cpp b/reg-lib/Platform.cpp
index 271273f4..19826418 100755
--- a/reg-lib/Platform.cpp
+++ b/reg-lib/Platform.cpp
@@ -1,6 +1,6 @@
 #include "Platform.h"
 #include "CpuKernelFactory.h"
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
 #include "CudaContext.hpp"
 #include "CudaF3dContent.h"
 #include "CudaComputeFactory.h"
@@ -9,7 +9,7 @@
 #include "CudaMeasureFactory.h"
 #include "_reg_optimiser_gpu.h"
 #endif
-#ifdef _USE_OPENCL
+#ifdef USE_OPENCL
 #include "ClContextSingleton.h"
 #include "ClComputeFactory.h"
 #include "ClContentCreatorFactory.h"
@@ -26,7 +26,7 @@ Platform::Platform(const PlatformType& platformTypeIn) {
         kernelFactory = new CpuKernelFactory();
         measureFactory = new MeasureFactory();
     }
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
     else if (platformType == PlatformType::Cuda) {
         platformName = "CUDA";
         SetGpuIdx(999);
@@ -36,7 +36,7 @@ Platform::Platform(const PlatformType& platformTypeIn) {
         measureFactory = new CudaMeasureFactory();
     }
 #endif
-#ifdef _USE_OPENCL
+#ifdef USE_OPENCL
     else if (platformType == PlatformType::OpenCl) {
         platformName = "OpenCL";
         SetGpuIdx(999);
@@ -71,7 +71,7 @@ void Platform::SetGpuIdx(unsigned gpuIdxIn) {
     if (platformType == PlatformType::Cpu) {
         gpuIdx = 999;
     }
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
     else if (platformType == PlatformType::Cuda) {
         CudaContext& cudaContext = CudaContext::GetInstance();
         if (gpuIdxIn != 999) {
@@ -80,7 +80,7 @@ void Platform::SetGpuIdx(unsigned gpuIdxIn) {
         }
     }
 #endif
-#ifdef _USE_OPENCL
+#ifdef USE_OPENCL
     else if (platformType == PlatformType::OpenCl) {
         ClContextSingleton& clContext = ClContextSingleton::GetInstance();
         if (gpuIdxIn != 999) {
@@ -138,7 +138,7 @@ reg_optimiser<Type>* Platform::CreateOptimiser(F3dContent& con,
             transformationGradientDataBw = (Type*)conBw->GetTransformationGradient()->data;
         }
     }
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
     else if (platformType == PlatformType::Cuda) {
         optimiser = dynamic_cast<reg_optimiser<Type>*>(useConjGradient ? new reg_conjugateGradient_gpu() : new reg_optimiser_gpu());
         controlPointGridData = (Type*)dynamic_cast<CudaF3dContent&>(con).GetControlPointGridCuda();
diff --git a/reg-lib/Platform.h b/reg-lib/Platform.h
index 5c7ed55f..42a0a823 100755
--- a/reg-lib/Platform.h
+++ b/reg-lib/Platform.h
@@ -10,10 +10,10 @@
 enum class PlatformType { Cpu, Cuda, OpenCl };
 constexpr PlatformType PlatformTypes[] = {
     PlatformType::Cpu,
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
     PlatformType::Cuda,
 #endif
-#ifdef _USE_OPENCL
+#ifdef USE_OPENCL
     PlatformType::OpenCl
 #endif
 };
@@ -43,13 +43,13 @@ class Platform {
                                          F3dContent *conBw = nullptr) const;
 
     static constexpr bool IsCudaEnabled() {
-#ifdef _USE_CUDA
+#ifdef USE_CUDA
         return true;
 #endif
         return false;
     }
     static constexpr bool IsOpenClEnabled() {
-#ifdef _USE_OPENCL
+#ifdef USE_OPENCL
         return true;
 #endif
         return false;
diff --git a/reg-lib/cpu/_reg_localTrans.cpp b/reg-lib/cpu/_reg_localTrans.cpp
index 15185c8a..6f95de7a 100755
--- a/reg-lib/cpu/_reg_localTrans.cpp
+++ b/reg-lib/cpu/_reg_localTrans.cpp
@@ -14,7 +14,7 @@
 #include "_reg_maths_eigen.h"
 
 #ifdef BUILD_TESTS
-#undef _USE_SSE
+#undef USE_SSE
 #endif
 
 /* *************************************************************** */
@@ -559,7 +559,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                                             int *mask,
                                             bool composition,
                                             bool bspline) {
-#if _USE_SSE
+#if USE_SSE
     union {
         __m128 m;
         float f[4];
@@ -596,13 +596,13 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
         DataType f[16] __attribute__((aligned(16)));
     } xyBasis;
 #endif // _WIN32
-#else // _USE_SSE
+#else // USE_SSE
     DataType xBasis[4];
     DataType yBasis[4];
     DataType xyBasis[16];
     DataType xControlPointCoordinates[16];
     DataType yControlPointCoordinates[16];
-#endif // _USE_SSE
+#endif // USE_SSE
 
     DataType *controlPointPtrX = static_cast<DataType*>(splineControlPoint->data);
     DataType *controlPointPtrY = &controlPointPtrX[NiftiImage::calcVoxelNumber(splineControlPoint, 2)];
@@ -659,7 +659,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                         yVoxel >= 0 && yVoxel <= deformationField->ny - 1) {
                         // The control point positions are extracted
                         if (oldXpre != xPre || oldYpre != yPre) {
-#ifdef _USE_SSE
+#ifdef USE_SSE
                             get_GridValues<DataType>(xPre,
                                                      yPre,
                                                      splineControlPoint,
@@ -669,7 +669,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                                                      yControlPointCoordinates.f,
                                                      false,  // no approximation
                                                      false); // not a displacement field
-#else // _USE_SSE
+#else // USE_SSE
                             get_GridValues<DataType>(xPre,
                                                      yPre,
                                                      splineControlPoint,
@@ -679,11 +679,11 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                                                      yControlPointCoordinates,
                                                      false,  // no approximation
                                                      false); // not a displacement field
-#endif // _USE_SSE
+#endif // USE_SSE
                             oldXpre = xPre;
                             oldYpre = yPre;
                         }
-#if _USE_SSE
+#if USE_SSE
                         coord = 0;
                         for (b = 0; b < 4; b++)
                             for (a = 0; a < 4; a++)
@@ -722,14 +722,14 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
         }
     } else { // starting deformation field is blank - !composition
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma  omp parallel for default(none) \
    shared(deformationField, gridVoxelSpacing, splineControlPoint, controlPointPtrX, \
    controlPointPtrY, mask, fieldPtrX, fieldPtrY, bspline) \
    private(x, a, xPre, yPre, oldXpre, oldYpre, index, xReal, yReal, basis, \
    val, xBasis, yBasis, tempCurrent, xyBasis, tempX, tempY, \
    xControlPointCoordinates, yControlPointCoordinates)
-#else // _USE_SSE
+#else // USE_SSE
 #pragma  omp parallel for default(none) \
    shared(deformationField, gridVoxelSpacing, splineControlPoint, controlPointPtrX, \
    controlPointPtrY, mask, fieldPtrX, fieldPtrY, bspline) \
@@ -753,7 +753,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                 if (basis < 0) basis = 0; // rounding error
                 if (bspline) get_BSplineBasisValues<DataType>(basis, xBasis);
                 else get_SplineBasisValues<DataType>(basis, xBasis);
-#if _USE_SSE
+#if USE_SSE
                 val.f[0] = static_cast<float>(xBasis[0]);
                 val.f[1] = static_cast<float>(xBasis[1]);
                 val.f[2] = static_cast<float>(xBasis[2]);
@@ -773,7 +773,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                 }
 #endif
                 if (oldXpre != xPre || oldYpre != yPre) {
-#ifdef _USE_SSE
+#ifdef USE_SSE
                     get_GridValues<DataType>(xPre,
                                              yPre,
                                              splineControlPoint,
@@ -783,7 +783,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                                              yControlPointCoordinates.f,
                                              false,  // no approximation
                                              false); // not a deformation field
-#else // _USE_SSE
+#else // USE_SSE
                     get_GridValues<DataType>(xPre,
                                              yPre,
                                              splineControlPoint,
@@ -793,7 +793,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                                              yControlPointCoordinates,
                                              false,  // no approximation
                                              false); // not a deformation field
-#endif // _USE_SSE
+#endif // USE_SSE
                     oldXpre = xPre;
                     oldYpre = yPre;
                 }
@@ -802,7 +802,7 @@ void reg_cubic_spline_getDeformationField2D(nifti_image *splineControlPoint,
                 yReal = 0;
 
                 if (mask[index] > -1) {
-#if _USE_SSE
+#if USE_SSE
                     tempX = _mm_set_ps1(0);
                     tempY = _mm_set_ps1(0);
                     //addition and multiplication of the 64 basis value and CP displacement for each axis
@@ -837,7 +837,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                             bool composition,
                                             bool bspline,
                                             bool forceNoLut = false) {
-#if _USE_SSE
+#if USE_SSE
     union {
         __m128 m;
         float f[4];
@@ -876,14 +876,14 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
         DataType f[16] __attribute__((aligned(16)));
     } zControlPointCoordinates;
 #endif // _WIN32
-#else // _USE_SSE
+#else // USE_SSE
     DataType temp[4];
     DataType zBasis[4];
     DataType xControlPointCoordinates[64];
     DataType yControlPointCoordinates[64];
     DataType zControlPointCoordinates[64];
     int coord;
-#endif // _USE_SSE
+#endif // USE_SSE
 
     const size_t splineControlPointVoxelNumber = NiftiImage::calcVoxelNumber(splineControlPoint, 3);
     DataType *controlPointPtrX = static_cast<DataType*>(splineControlPoint->data);
@@ -906,7 +906,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
         if (splineControlPoint->sform_code > 0)
             referenceMatrix_real_to_voxel = splineControlPoint->sto_ijk;
         else referenceMatrix_real_to_voxel = splineControlPoint->qto_ijk;
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #ifdef _WIN32
         __declspec(align(16)) DataType xBasis[4];
         __declspec(align(16)) DataType yBasis[4];
@@ -914,14 +914,14 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
         DataType xBasis[4] __attribute__((aligned(16)));
         DataType yBasis[4] __attribute__((aligned(16)));
 #endif
-#else // _USE_SSE
+#else // USE_SSE
         DataType xBasis[4], yBasis[4];
-#endif // _USE_SSE
+#endif // USE_SSE
 
         DataType voxel[3];
 
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma omp parallel for default(none) \
    private(x, y, b, c, oldPreX, oldPreY, oldPreZ, xPre, yPre, zPre, real, \
    index, voxel, basis, xBasis, yBasis, zBasis, xControlPointCoordinates, \
@@ -939,7 +939,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
    shared(deformationField, fieldPtrX, fieldPtrY, fieldPtrZ, referenceMatrix_real_to_voxel, \
    bspline, controlPointPtrX, controlPointPtrY, controlPointPtrZ, \
    splineControlPoint, mask)
-#endif // _USE_SSE
+#endif // USE_SSE
 #endif // _OPENMP
         for (z = 0; z < deformationField->nz; z++) {
             index = z * deformationField->nx * deformationField->ny;
@@ -990,7 +990,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
 
                         // The control point positions are extracted
                         if (xPre != oldPreX || yPre != oldPreY || zPre != oldPreZ) {
-#ifdef _USE_SSE
+#ifdef USE_SSE
                             get_GridValues<DataType>(xPre,
                                                      yPre,
                                                      zPre,
@@ -1003,7 +1003,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                      zControlPointCoordinates.f,
                                                      false,  // no approximation
                                                      false); // not a deformation field
-#else // _USE_SSE
+#else // USE_SSE
                             get_GridValues<DataType>(xPre,
                                                      yPre,
                                                      zPre,
@@ -1016,13 +1016,13 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                      zControlPointCoordinates,
                                                      false,  // no approximation
                                                      false); // not a deformation field
-#endif // _USE_SSE
+#endif // USE_SSE
                             oldPreX = xPre;
                             oldPreY = yPre;
                             oldPreZ = zPre;
                         }
 
-#if _USE_SSE
+#if USE_SSE
                         tempX = _mm_set_ps1(0);
                         tempY = _mm_set_ps1(0);
                         tempZ = _mm_set_ps1(0);
@@ -1083,7 +1083,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
         gridVoxelSpacing[1] = splineControlPoint->dy / deformationField->dy;
         gridVoxelSpacing[2] = splineControlPoint->dz / deformationField->dz;
 
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #ifdef _WIN32
         union u1 {
             __m128 m[4];
@@ -1103,9 +1103,9 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
             DataType f[64] __attribute__((aligned(16)));
         } xyzBasis;
 #endif // _WIN32
-#else // _USE_SSE
+#else // USE_SSE
         DataType yzBasis[16], xyzBasis[64];
-#endif // _USE_SSE
+#endif // USE_SSE
 
         // Assess if lookup table can be used
         if (gridVoxelSpacing[0] == 5. && gridVoxelSpacing[0] == 5. && gridVoxelSpacing[0] == 5. && forceNoLut == false) {
@@ -1114,15 +1114,15 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
             // Compute and store all required coefficients
             int coeff_index;
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma omp parallel for default(none) \
     private(x, y, a, coeff_index, basis, zBasis, temp, val, tempCurrent, yzBasis) \
     shared(coefficients, bspline)
-#else //  _USE_SSE
+#else //  USE_SSE
 #pragma omp parallel for default(none) \
     private(x, y, a, coeff_index, basis, zBasis, temp, yzBasis, coord) \
     shared(coefficients, bspline)
-#endif // _USE_SSE
+#endif // USE_SSE
 #endif // _OPENMP
             for (z = 0; z < 5; ++z) {
                 coeff_index = z * 5 * 5 * 64;
@@ -1133,7 +1133,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                     basis = static_cast<DataType>(y) / 5.f;
                     if (bspline) get_BSplineBasisValues<DataType>(basis, temp);
                     else get_SplineBasisValues<DataType>(basis, temp);
-#if _USE_SSE
+#if USE_SSE
                     val.f[0] = static_cast<float>(temp[0]);
                     val.f[1] = static_cast<float>(temp[1]);
                     val.f[2] = static_cast<float>(temp[2]);
@@ -1157,7 +1157,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                         basis = static_cast<DataType>(x) / 5.f;
                         if (bspline) get_BSplineBasisValues<DataType>(basis, temp);
                         else get_SplineBasisValues<DataType>(basis, temp);
-#if _USE_SSE
+#if USE_SSE
                         val.f[0] = static_cast<float>(temp[0]);
                         val.f[1] = static_cast<float>(temp[1]);
                         val.f[2] = static_cast<float>(temp[2]);
@@ -1184,11 +1184,11 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
             } // z
 
             // Loop over block of 5x5x5 voxels
-#if _USE_SSE
+#if USE_SSE
             int coord;
 #endif // USE_SSE
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma omp parallel for default(none) \
    private(x, y, z, a, b, c, xPre, yPre, real, \
    index, coeff_index, coord, tempX, tempY, tempZ, val,\
@@ -1196,7 +1196,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
    shared(deformationField, fieldPtrX, fieldPtrY, fieldPtrZ, splineControlPoint, mask, \
    gridVoxelSpacing, bspline, controlPointPtrX, controlPointPtrY, controlPointPtrZ, \
    coefficients)
-#else //  _USE_SSE
+#else //  USE_SSE
 #pragma omp parallel for default(none) \
    private(x, y, z, a, b, c, xPre, yPre, real, \
    index, coeff_index, coord, basis, \
@@ -1204,12 +1204,12 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
    shared(deformationField, fieldPtrX, fieldPtrY, fieldPtrZ, splineControlPoint, mask, \
    gridVoxelSpacing, bspline, controlPointPtrX, controlPointPtrY, controlPointPtrZ, \
    coefficients)
-#endif // _USE_SSE
+#endif // USE_SSE
 #endif // _OPENMP
             for (zPre = 0; zPre < splineControlPoint->nz - 3; zPre++) {
                 for (yPre = 0; yPre < splineControlPoint->ny - 3; yPre++) {
                     for (xPre = 0; xPre < splineControlPoint->nx - 3; xPre++) {
-#if _USE_SSE
+#if USE_SSE
                         get_GridValues<DataType>(xPre,
                                                  yPre,
                                                  zPre,
@@ -1222,7 +1222,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                  zControlPointCoordinates.f,
                                                  false,  // no approximation
                                                  false); // not a deformation field
-#else // _USE_SSE
+#else // USE_SSE
                         get_GridValues<DataType>(xPre,
                                                  yPre,
                                                  zPre,
@@ -1235,7 +1235,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                  zControlPointCoordinates,
                                                  false,  // no approximation
                                                  false); // not a deformation field
-#endif // _USE_SSE
+#endif // USE_SSE
                         coeff_index = 0;
                         for (c = 0; c < 5; ++c) {
                             z = zPre * 5 + c;
@@ -1247,7 +1247,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                         for (a = 0; a < 5; ++a) {
                                             x = xPre * 5 + a;
                                             if (x<deformationField->nx && mask[index]>-1) {
-#if _USE_SSE
+#if USE_SSE
                                                 tempX = _mm_set_ps1(0);
                                                 tempY = _mm_set_ps1(0);
                                                 tempZ = _mm_set_ps1(0);
@@ -1276,7 +1276,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                 val.m = tempZ;
                                                 real[2] = val.f[0] + val.f[1] + val.f[2] + val.f[3];
 #endif
-#else // _USE_SSE
+#else // USE_SSE
                                                 real[0] = real[1] = real[2] = 0;
                                                 for (coord = 0; coord < 64; ++coord) {
                                                     basis = coefficients[coeff_index++];
@@ -1284,7 +1284,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                     real[1] += yControlPointCoordinates[coord] * basis;
                                                     real[2] += zControlPointCoordinates[coord] * basis;
                                                 }
-#endif // _USE_SSE
+#endif // USE_SSE
                                                 fieldPtrX[index] = real[0];
                                                 fieldPtrY[index] = real[1];
                                                 fieldPtrZ[index] = real[2];
@@ -1304,7 +1304,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
             free(coefficients);
         } else { // if spacings!=5 voxels
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma omp parallel for default(none) \
     private(x, y, a, xPre, yPre, zPre, real, \
     index, basis, xyzBasis, yzBasis, zBasis, temp, xControlPointCoordinates, \
@@ -1313,14 +1313,14 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
     temp_basis_sse, basis_sse, val, tempCurrent) \
     shared(deformationField, fieldPtrX, fieldPtrY, fieldPtrZ, splineControlPoint, mask, \
     gridVoxelSpacing, bspline, controlPointPtrX, controlPointPtrY, controlPointPtrZ)
-#else //  _USE_SSE
+#else //  USE_SSE
 #pragma omp parallel for default(none) \
     private(x, y, a, xPre, yPre, zPre, real, \
     index, basis, xyzBasis, yzBasis, zBasis, temp, xControlPointCoordinates, \
     yControlPointCoordinates, zControlPointCoordinates, oldBasis, coord) \
     shared(deformationField, fieldPtrX, fieldPtrY, fieldPtrZ, splineControlPoint, mask, \
     gridVoxelSpacing, bspline, controlPointPtrX, controlPointPtrY, controlPointPtrZ)
-#endif // _USE_SSE
+#endif // USE_SSE
 #endif // _OPENMP
             for (z = 0; z < deformationField->nz; z++) {
                 index = z * deformationField->nx * deformationField->ny;
@@ -1338,7 +1338,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                     if (basis < 0) basis = 0; //rounding error
                     if (bspline) get_BSplineBasisValues<DataType>(basis, temp);
                     else get_SplineBasisValues<DataType>(basis, temp);
-#if _USE_SSE
+#if USE_SSE
                     val.f[0] = static_cast<float>(temp[0]);
                     val.f[1] = static_cast<float>(temp[1]);
                     val.f[2] = static_cast<float>(temp[2]);
@@ -1363,7 +1363,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                         if (basis < 0) basis = 0; //rounding error
                         if (bspline) get_BSplineBasisValues<DataType>(basis, temp);
                         else get_SplineBasisValues<DataType>(basis, temp);
-#if _USE_SSE
+#if USE_SSE
                         val.f[0] = static_cast<float>(temp[0]);
                         val.f[1] = static_cast<float>(temp[1]);
                         val.f[2] = static_cast<float>(temp[2]);
@@ -1383,7 +1383,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                         }
 #endif
                         if (basis <= oldBasis || x == 0) {
-#ifdef _USE_SSE
+#ifdef USE_SSE
                             get_GridValues<DataType>(xPre,
                                                      yPre,
                                                      zPre,
@@ -1396,7 +1396,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                      zControlPointCoordinates.f,
                                                      false,  // no approximation
                                                      false); // not a deformation field
-#else // _USE_SSE
+#else // USE_SSE
                             get_GridValues<DataType>(xPre,
                                                      yPre,
                                                      zPre,
@@ -1409,7 +1409,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                                                      zControlPointCoordinates,
                                                      false,  // no approximation
                                                      false); // not a deformation field
-#endif // _USE_SSE
+#endif // USE_SSE
                         }
                         oldBasis = basis;
 
@@ -1418,7 +1418,7 @@ void reg_cubic_spline_getDeformationField3D(nifti_image *splineControlPoint,
                         real[2] = 0;
 
                         if (mask[index] > -1) {
-#if _USE_SSE
+#if USE_SSE
                             tempX = _mm_set_ps1(0);
                             tempY = _mm_set_ps1(0);
                             tempZ = _mm_set_ps1(0);
@@ -1463,7 +1463,7 @@ void reg_spline_getDeformationField(nifti_image *splineControlPoint,
     if (splineControlPoint->datatype != deformationField->datatype)
         NR_FATAL_ERROR("The spline control point image and the deformation field image are expected to be of the same type");
 
-#if _USE_SSE
+#if USE_SSE
     if (splineControlPoint->datatype != NIFTI_TYPE_FLOAT32)
         NR_FATAL_ERROR("SSE computation has only been implemented for single precision");
 #endif
@@ -3056,12 +3056,12 @@ void reg_spline_cppComposition_2D(nifti_image *grid1,
                                   bool bspline) {
     // REMINDER Grid2(x)=Grid1(Grid2(x))
 
-#if _USE_SSE
+#if USE_SSE
     union {
         __m128 m;
         float f[4];
     } val;
-#endif // _USE_SSE
+#endif // USE_SSE
 
     DataType *outCPPPtrX = static_cast<DataType*>(grid2->data);
     DataType *outCPPPtrY = &outCPPPtrX[NiftiImage::calcVoxelNumber(grid2, 2)];
@@ -3074,18 +3074,18 @@ void reg_spline_cppComposition_2D(nifti_image *grid1,
 #ifdef _WIN32
     __declspec(align(16)) DataType xBasis[4];
     __declspec(align(16)) DataType yBasis[4];
-#if _USE_SSE
+#if USE_SSE
     __declspec(align(16)) DataType xyBasis[16];
-#endif  //_USE_SSE
+#endif  //USE_SSE
 
     __declspec(align(16)) DataType xControlPointCoordinates[16];
     __declspec(align(16)) DataType yControlPointCoordinates[16];
 #else // _WIN32
     DataType xBasis[4] __attribute__((aligned(16)));
     DataType yBasis[4] __attribute__((aligned(16)));
-#if _USE_SSE
+#if USE_SSE
     DataType xyBasis[16] __attribute__((aligned(16)));
-#endif  //_USE_SSE
+#endif  //USE_SSE
 
     DataType xControlPointCoordinates[16] __attribute__((aligned(16)));
     DataType yControlPointCoordinates[16] __attribute__((aligned(16)));
@@ -3153,7 +3153,7 @@ void reg_spline_cppComposition_2D(nifti_image *grid1,
                                      displacement1); // displacement field?
             xReal = 0;
             yReal = 0;
-#if _USE_SSE
+#if USE_SSE
             coord = 0;
             for (unsigned b = 0; b < 4; b++) {
                 for (unsigned a = 0; a < 4; a++) {
@@ -3206,7 +3206,7 @@ void reg_spline_cppComposition_3D(nifti_image *grid1,
                                   bool displacement2,
                                   bool bspline) {
     // REMINDER Grid2(x)=Grid1(Grid2(x))
-#if _USE_SSE
+#if USE_SSE
     union {
         __m128 m;
         float f[4];
@@ -3272,7 +3272,7 @@ void reg_spline_cppComposition_3D(nifti_image *grid1,
     else matrix_voxel_to_real2 = &grid2->qto_xyz;
 
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma omp parallel for default(none) \
    shared(grid1, grid2, displacement1, displacement2, matrix_voxel_to_real2, matrix_real_to_voxel1, \
    outCPPPtrX, outCPPPtrY, outCPPPtrZ, controlPointPtrX, controlPointPtrY, controlPointPtrZ, bspline) \
@@ -3380,7 +3380,7 @@ void reg_spline_cppComposition_3D(nifti_image *grid1,
                 xReal = 0;
                 yReal = 0;
                 zReal = 0;
-#if _USE_SSE
+#if USE_SSE
                 val.f[0] = static_cast<float>(xBasis[0]);
                 val.f[1] = static_cast<float>(xBasis[1]);
                 val.f[2] = static_cast<float>(xBasis[2]);
@@ -3450,7 +3450,7 @@ int reg_spline_cppComposition(nifti_image *grid1,
     if (grid1->datatype != grid2->datatype)
         NR_FATAL_ERROR("Both input images are expected to have the same data type");
 
-#if _USE_SSE
+#if USE_SSE
     if (grid1->datatype != NIFTI_TYPE_FLOAT32)
         NR_FATAL_ERROR("SSE computation has only been implemented for single precision");
 #endif
diff --git a/reg-lib/cpu/_reg_localTrans_jac.cpp b/reg-lib/cpu/_reg_localTrans_jac.cpp
index 26678dde..75c0b6ee 100755
--- a/reg-lib/cpu/_reg_localTrans_jac.cpp
+++ b/reg-lib/cpu/_reg_localTrans_jac.cpp
@@ -12,7 +12,7 @@
 
 #include "_reg_localTrans_jac.h"
 
-#define _USE_SQUARE_LOG_JAC
+#define USE_SQUARE_LOG_JAC
 
 /* *************************************************************** */
 /* *************************************************************** */
@@ -643,7 +643,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
       // Allocate variables that are used in both scenarii
       int pre[3], oldPre[3], incr0;
       DataType basis, xBasis[4], xFirst[4], yBasis[4], yFirst[4], zBasis[4], zFirst[4];
-#if _USE_SSE
+#if USE_SSE
       union
       {
          __m128 m;
@@ -805,7 +805,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                   basis = gridCoord[2] - pre[2];
                   get_BSplineBasisValues<DataType>(basis, zBasis, zFirst);
                   // Compute the 64 basis values and the corresponding derivatives
-#if _USE_SSE
+#if USE_SSE
                   val.f[0]=yBasis[0];
                   val.f[1]=yBasis[1];
                   val.f[2]=yBasis[2];
@@ -862,7 +862,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                   // Fetch the required coefficients
                   if(oldPre[0]!=pre[0] || oldPre[1]!=pre[1] || oldPre[2]!=pre[2])
                   {
-#ifdef _USE_SSE
+#ifdef USE_SSE
                      get_GridValues<DataType>(pre[0]-1,
                            pre[1]-1,
                            pre[2]-1,
@@ -876,7 +876,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                            false, // no approx
                            false // not disp
                            );
-#else // _USE_SSE
+#else // USE_SSE
                      get_GridValues<DataType>(pre[0]-1,
                            pre[1]-1,
                            pre[2]-1,
@@ -890,13 +890,13 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                            false, // no approx
                            false // not disp
                            );
-#endif // _USE_SSE
+#endif // USE_SSE
                      oldPre[0]=pre[0];
                      oldPre[1]=pre[1];
                      oldPre[2]=pre[2];
                   }
                   // Compute the Jacobian matrix
-#if _USE_SSE
+#if USE_SSE
                   tempX_x =  _mm_set_ps1(0);
                   tempX_y =  _mm_set_ps1(0);
                   tempX_z =  _mm_set_ps1(0);
@@ -973,7 +973,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
       {
          // The grid is assumed to be aligned with the reference image
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma omp parallel for default(none) \
    shared(referenceImage, gridVoxelSpacing, splineControlPoint, \
    coeffPtrX, coeffPtrY, coeffPtrZ,reorientation, JacobianMatrices, \
@@ -1015,7 +1015,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                if(basis<0) basis=0; //rounding error
                get_BSplineBasisValues<DataType>(basis, yBasis, yFirst);
 
-#if _USE_SSE
+#if USE_SSE
                val.f[0]=yBasis[0];
                val.f[1]=yBasis[1];
                val.f[2]=yBasis[2];
@@ -1055,7 +1055,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                   if(basis<0) basis=0; //rounding error
                   get_BSplineBasisValues<DataType>(basis, xBasis, xFirst);
 
-#if _USE_SSE
+#if USE_SSE
                   val.f[0]=xBasis[0];
                   val.f[1]=xBasis[1];
                   val.f[2]=xBasis[2];
@@ -1091,7 +1091,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
 
                   if(oldPre[0]!=pre[0] || oldPre[1]!=pre[1] || oldPre[2]!=pre[2])
                   {
-#ifdef _USE_SSE
+#ifdef USE_SSE
                      get_GridValues<DataType>(pre[0],
                            pre[1],
                            pre[2],
@@ -1105,7 +1105,7 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                            false, // no approx
                            false // not disp
                            );
-#else // _USE_SSE
+#else // USE_SSE
                      get_GridValues<DataType>(pre[0],
                            pre[1],
                            pre[2],
@@ -1119,12 +1119,12 @@ void reg_cubic_spline_jacobian3D(nifti_image *splineControlPoint,
                            false, // no approx
                            false // not disp
                            );
-#endif // _USE_SSE
+#endif // USE_SSE
                      oldPre[0]=pre[0];
                      oldPre[1]=pre[1];
                      oldPre[2]=pre[2];
                   }
-#if _USE_SSE
+#if USE_SSE
                   tempX_x =  _mm_set_ps1(0);
                   tempX_y =  _mm_set_ps1(0);
                   tempX_z =  _mm_set_ps1(0);
@@ -1278,7 +1278,7 @@ double reg_spline_getJacobianPenaltyTerm(nifti_image *splineControlPoint,
          for(size_t i=0; i<detNumber; ++i)
          {
             double logDet = log(jacDetPtr[i]);
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
             penaltySum += logDet * logDet;
 #else
             penaltySum += fasb(logDet);
@@ -1292,7 +1292,7 @@ double reg_spline_getJacobianPenaltyTerm(nifti_image *splineControlPoint,
          for(size_t i=0; i<detNumber; ++i)
          {
             double logDet = log(jacDetPtr[i]);
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
             penaltySum += logDet * logDet;
 #else
             penaltySum += fasb(logDet);
@@ -1410,7 +1410,7 @@ void reg_spline_jacobianDetGradient2D(nifti_image *splineControlPoint,
                         if(detJac>0)
                         {
                            jacobianMatrix = jacobianMatrices[jacIndex];
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
                            detJac = 2.0*log(detJac) / detJac;
 #else
                            detJac = (log(detJac)>0?1.0:-1.0) / detJac;
@@ -1513,7 +1513,7 @@ void reg_spline_jacobianDetGradient2D(nifti_image *splineControlPoint,
                               basisValues[1] = xBasis * yFirst ;
 
                               jacobianMatrix = jacobianMatrices[jacIndex];
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
                               detJac= 2.0*log(detJac) / detJac;
 #else
                               detJac = (log(detJac)>0?1.0:-1.0) / detJac;
@@ -1661,7 +1661,7 @@ void reg_spline_jacobianDetGradient3D(nifti_image *splineControlPoint,
                                  if(detJac>0)
                                  {
                                     jacobianMatrix = jacobianMatrices[jacIndex];
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
                                     detJac = 2.0*log(detJac) / detJac;
 #else
                                     detJac = (log(detJac)>0?1.0:-1.0) / detJac;
@@ -1787,7 +1787,7 @@ void reg_spline_jacobianDetGradient3D(nifti_image *splineControlPoint,
                                        basisValues[2] = xBasis * yBasis * zFirst ;
 
                                        jacobianMatrix = jacobianMatrices[jacIndex];
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
                                        detJac= 2.0*log(detJac) / detJac;
 #else
                                        detJac = (log(detJac)>0?1.0:-1.0) / detJac;
@@ -1931,7 +1931,7 @@ double reg_spline_correctFolding2D(nifti_image *splineControlPoint,
    for(i=0; i< jacobianNumber; i++)
    {
       logDet = log(jacobianDeterminant[i]);
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
       penaltyTerm += logDet*logDet;
 #else
       penaltyTerm +=  fabs(log(logDet));
@@ -2180,7 +2180,7 @@ double reg_spline_correctFolding3D(nifti_image *splineControlPoint,
    for(i=0; i< jacobianNumber; i++)
    {
       logDet = log(jacobianDeterminant[i]);
-#ifdef _USE_SQUARE_LOG_JAC
+#ifdef USE_SQUARE_LOG_JAC
       penaltyTerm += logDet*logDet;
 #else
       penaltyTerm +=  fabs(log(logDet));
diff --git a/reg-lib/cpu/_reg_maths.h b/reg-lib/cpu/_reg_maths.h
index 93151883..c77e18fd 100644
--- a/reg-lib/cpu/_reg_maths.h
+++ b/reg-lib/cpu/_reg_maths.h
@@ -20,7 +20,7 @@
 #include <omp.h>
 #endif
 
-#if _USE_SSE
+#if USE_SSE
 #include <emmintrin.h>
 #include <xmmintrin.h>
 #ifdef __SSE3__
diff --git a/reg-lib/cpu/_reg_tools.cpp b/reg-lib/cpu/_reg_tools.cpp
index 93a0a76c..91a85e3a 100755
--- a/reg-lib/cpu/_reg_tools.cpp
+++ b/reg-lib/cpu/_reg_tools.cpp
@@ -954,7 +954,7 @@ void reg_tools_kernelConvolution(nifti_image *image,
                         double bufferIntensityCur = 0;
                         double bufferDensityCur = 0;
 
-#ifdef _USE_SSE
+#ifdef USE_SSE
                         union {
                             __m128 m;
                             float f[4];
@@ -963,7 +963,7 @@ void reg_tools_kernelConvolution(nifti_image *image,
 #endif
 
 #ifdef _OPENMP
-#ifdef _USE_SSE
+#ifdef USE_SSE
 #pragma omp parallel for default(none) \
    shared(imageDims, intensityPtr, densityPtr, radius, kernel, lineOffset, n, planeNumber, kernelSum) \
    private(realIndex, currentIntensityPtr, currentDensityPtr, lineIndex, bufferIntensity, \
@@ -1015,7 +1015,7 @@ void reg_tools_kernelConvolution(nifti_image *image,
                                     if (shiftPst > imageDims[n]) shiftPst = imageDims[n];
                                     // Set the current values to zero
                                     // Increment the current value by performing the weighted sum
-#ifdef _USE_SSE
+#ifdef USE_SSE
                                     intensity_sum_sse.m = _mm_set_ps1(0);
                                     density_sum_sse.m = _mm_set_ps1(0);
                                     k = shiftPre;
diff --git a/reg-test/reg_test_be.cpp b/reg-test/reg_test_be.cpp
index 445d3959..afe18f83 100644
--- a/reg-test/reg_test_be.cpp
+++ b/reg-test/reg_test_be.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_composeField.cpp b/reg-test/reg_test_composeField.cpp
index 6bd7662e..49550c77 100644
--- a/reg-test/reg_test_composeField.cpp
+++ b/reg-test/reg_test_composeField.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_conjugateGradient.cpp b/reg-test/reg_test_conjugateGradient.cpp
index bb2d4e63..644eb49b 100644
--- a/reg-test/reg_test_conjugateGradient.cpp
+++ b/reg-test/reg_test_conjugateGradient.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_getDeformationField.cpp b/reg-test/reg_test_getDeformationField.cpp
index c49a1a24..b213f3fc 100644
--- a/reg-test/reg_test_getDeformationField.cpp
+++ b/reg-test/reg_test_getDeformationField.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_imageGradient.cpp b/reg-test/reg_test_imageGradient.cpp
index 8689954a..1b243132 100644
--- a/reg-test/reg_test_imageGradient.cpp
+++ b/reg-test/reg_test_imageGradient.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_interpolation.cpp b/reg-test/reg_test_interpolation.cpp
index 3de5aae3..b3d05830 100644
--- a/reg-test/reg_test_interpolation.cpp
+++ b/reg-test/reg_test_interpolation.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_lncc.cpp b/reg-test/reg_test_lncc.cpp
index 6e45e6d1..0355aa84 100644
--- a/reg-test/reg_test_lncc.cpp
+++ b/reg-test/reg_test_lncc.cpp
@@ -1,6 +1,6 @@
 // OpenCL and CUDA are not supported for this test yet
-#undef _USE_OPENCL
-#undef _USE_CUDA
+#undef USE_OPENCL
+#undef USE_CUDA
 
 #include "reg_test_common.h"
 #include "_reg_lncc.h"
diff --git a/reg-test/reg_test_nmi.cpp b/reg-test/reg_test_nmi.cpp
index 1b61ac39..21847f10 100644
--- a/reg-test/reg_test_nmi.cpp
+++ b/reg-test/reg_test_nmi.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test yet
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 #include "_reg_tools.h"
diff --git a/reg-test/reg_test_nmi_gradient.cpp b/reg-test/reg_test_nmi_gradient.cpp
index f19ac9bd..95283b0f 100644
--- a/reg-test/reg_test_nmi_gradient.cpp
+++ b/reg-test/reg_test_nmi_gradient.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test yet
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_normaliseGradient.cpp b/reg-test/reg_test_normaliseGradient.cpp
index d56cd356..53c6f40e 100644
--- a/reg-test/reg_test_normaliseGradient.cpp
+++ b/reg-test/reg_test_normaliseGradient.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_regr_getDeformationField.cpp b/reg-test/reg_test_regr_getDeformationField.cpp
index d5c0a8de..62955c0b 100644
--- a/reg-test/reg_test_regr_getDeformationField.cpp
+++ b/reg-test/reg_test_regr_getDeformationField.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"
 
diff --git a/reg-test/reg_test_voxelCentricToNodeCentric.cpp b/reg-test/reg_test_voxelCentricToNodeCentric.cpp
index 47f56f90..3339cbbc 100644
--- a/reg-test/reg_test_voxelCentricToNodeCentric.cpp
+++ b/reg-test/reg_test_voxelCentricToNodeCentric.cpp
@@ -1,5 +1,5 @@
 // OpenCL is not supported for this test
-#undef _USE_OPENCL
+#undef USE_OPENCL
 
 #include "reg_test_common.h"