diff --git a/examples/my_nodes/dp_node_ants_apply.m b/examples/my_nodes/dp_node_ants_apply.m
new file mode 100644
index 0000000..42cf348
--- /dev/null
+++ b/examples/my_nodes/dp_node_ants_apply.m
@@ -0,0 +1,45 @@
+classdef dp_node_ants_apply < dp_node
+    % Wrapper around antsApplyTransforms onto a given reference grid
+    % input.nii_fn   : image to be transformed (e.g. atlas labels in MNI)
+    % input.ref_fn   : reference grid / output space (e.g. trimmed FA)
+    % input.affine_fn, input.invwarp_fn : from dp_node_ants_reg_quick
+    % output.nii_fn  : resampled image in ref_fn space
+
+    properties
+        options = '';       % e.g. ' -n NearestNeighbor' for labels
+        use_inverse = 1;    % default = inverse chain
+    end
+
+    methods
+
+        function obj = dp_node_ants_apply(options)
+            if (nargin>0), obj.options = options; end
+            obj.output_test = {'nii_fn'};
+        end
+
+        function output = i2o(obj, input)
+            output.nii_fn = dp.new_fn(input.op, input.nii_fn, '_ants');
+            output.ref_fn = input.ref_fn;
+        end
+
+        function output = execute(obj, input, output)
+            msf_mkdir(fileparts(output.nii_fn));
+            if obj.use_inverse
+                % apply inverse chain: template→subject (fixed → moving)
+                t = sprintf('-t "[%s,1]" -t "%s" ', input.affine_fn, input.invwarp_fn);
+            else
+                % apply forward chain: subject→template (moving → fixed)
+                t = sprintf('-t "%s" -t "%s" ', input.warp_fn, input.affine_fn);
+            end
+
+            if isempty(obj.options), obj.options = ' -n NearestNeighbor'; end
+
+            % cmd using composed transform chain "t" (forward or inverse, depending on use_inverse)
+            cmd = sprintf('antsApplyTransforms -d 3 -i "%s" -r "%s" %s -o "%s"%s', ...
+                input.nii_fn, input.ref_fn, t, output.nii_fn, obj.options);
+
+
+            system(cmd);
+        end
+    end
+end
diff --git a/examples/my_nodes/dp_node_ants_reg_quick.m b/examples/my_nodes/dp_node_ants_reg_quick.m
new file mode 100644
index 0000000..c30f7be
--- /dev/null
+++ b/examples/my_nodes/dp_node_ants_reg_quick.m
@@ -0,0 +1,64 @@
+classdef dp_node_ants_reg_quick < dp_node
+    % antsRegistrationSyN.sh (moving = input.nii_fn, fixed = input.target_fn)
+    % Outputs affine + forward/inverse warps in <input.op>
+
+    properties
+        options = '';   % include threads and preset here, e.g. ' ITK_GLOBAL_DEFAULT_NUMBER_OF_THREADS=16 -t s'
+    end
+
+    methods
+
+        function obj = dp_node_ants_reg_quick(options)
+            if (nargin>0), obj.options = options; end
+            obj.output_test = {'affine_fn','warp_fn','invwarp_fn','warped_fn'};
+
+        end
+
+        function output = i2o(obj, input)
+            tmp = msf_fn_append(input.nii_fn, '_mni_');
+            prefix = msf_fn_new_path(input.op, tmp);
+            [pdir,pfx,~] = msf_fileparts(prefix);
+            prefix = fullfile(pdir, pfx);                     % ANTs prefix base (no ext)
+
+            output.prefix      = prefix;
+            output.target_fn   = input.target_fn;
+            output.affine_fn   = [prefix '0GenericAffine.mat'];
+            output.warp_fn     = [prefix '1Warp.nii.gz'];
+            output.invwarp_fn  = [prefix '1InverseWarp.nii.gz'];
+            output.warped_fn   = [prefix 'Warped.nii.gz'];    % QC, remains in op
+        end
+
+
+        function output = execute(obj, input, output)
+            % set threads and build the command
+            thr  = feature('numcores');
+            envp = sprintf('ITK_GLOBAL_DEFAULT_NUMBER_OF_THREADS=%d', thr);
+
+            % if user didn’t provide flags, default to symmetric registration (-t s)
+            if isempty(strtrim(obj.options))
+                flags = ' -t s';
+            else
+                flags = [' ' strtrim(obj.options)];
+            end
+
+            msf_mkdir(fileparts(output.warped_fn));
+
+
+            % full SyN variant
+            % cmd = sprintf('%s antsRegistrationSyN.sh -d 3 -f "%s" -m "%s" -o "%s"%s', ...
+            %     envp, input.target_fn, input.nii_fn, output.prefix, flags);
+
+            % quick SyN variant 
+            cmd = sprintf('%s antsRegistrationSyNQuick.sh -d 3 -f "%s" -m "%s" -o "%s"%s', ...
+                envp, input.target_fn, input.nii_fn, output.prefix, flags);
+
+
+            % Always save ANTs output to log in the same folder
+            log_fn = fullfile(fileparts(output.warped_fn), 'ants_registration.log');
+            cmd = [cmd, sprintf(' > "%s" 2>&1', log_fn)];
+
+            system(cmd);
+        end
+
+    end
+end
diff --git a/examples/my_nodes/dp_node_atlas_crop_to_mask.m b/examples/my_nodes/dp_node_atlas_crop_to_mask.m
new file mode 100644
index 0000000..9e6bc2b
--- /dev/null
+++ b/examples/my_nodes/dp_node_atlas_crop_to_mask.m
@@ -0,0 +1,53 @@
+classdef dp_node_atlas_crop_to_mask < dp_node
+    % Crop an atlas/label map with a provided mask (same grid).
+    %
+    % Inputs (from upstream):
+    %   nii_fn   - atlas / label image to crop (e.g. warped JHU labels)
+    %   mask_fn  - mask in the same space (e.g. trimmed FA brain mask)
+    %   op       - optional output directory
+    %
+    % Output:
+    %   nii_fn   - cropped atlas, saved as <op>/<basename>_crop.nii.gz
+
+    properties
+        suffix = '_crop';
+    end
+
+    methods
+        function obj = dp_node_atlas_crop_to_mask(suffix)
+            if nargin > 0
+                obj.suffix = suffix;
+            end
+            obj.output_test = {'nii_fn'};
+        end
+
+        function output = i2o(obj, input)
+            % Decide output folder
+            if isfield(input, 'op') && ~isempty(input.op)
+                op = input.op;
+            else
+                [op, ~, ~] = msf_fileparts(input.nii_fn);
+            end
+
+            % Use dp.new_fn to append suffix before extension
+            output.nii_fn = dp.new_fn(op, input.nii_fn, obj.suffix);
+        end
+
+        function output = execute(obj, input, output)
+            % Read atlas/labels and mask
+            [A, Ah] = mdm_nii_read(input.nii_fn);
+            B       = mdm_nii_read(input.mask_fn);
+
+            % Basic safety
+            if ~isequal(size(A), size(B))
+                error('dp_node_atlas_crop_to_mask: atlas/mask size mismatch');
+            end
+
+            msf_mkdir(fileparts(output.nii_fn));
+
+            % Apply crop
+            mdm_nii_write(A .* (B > 0), output.nii_fn, Ah);
+
+        end
+    end
+end
diff --git a/examples/my_nodes/dp_node_segm_jhu.m b/examples/my_nodes/dp_node_segm_jhu.m
new file mode 100644
index 0000000..80cdfab
--- /dev/null
+++ b/examples/my_nodes/dp_node_segm_jhu.m
@@ -0,0 +1,71 @@
+classdef dp_node_segm_jhu < dp_node_segm
+
+    methods (Hidden)
+        function [labels, ids] = segm_info(obj)
+
+            % /usr/local/fsl/data/atlases/JHU-labels.xml
+
+            txt = { ...
+                '0: Unclassified', ...       % Unclassified
+                '1: MCP', ...        % Middle cerebellar peduncle
+                '2: PCT', ...        % Pontine crossing tract (a part of MCP)
+                '3: GCC', ...        % Genu of corpus callosum
+                '4: CCB', ...        % Body of corpus callosum
+                '5: SCC', ...        % Splenium of corpus callosum
+                '6: FX', ...         % Fornix (column and body of fornix)
+                '7: CST_R', ...      % Corticospinal tract R
+                '8: CST_L', ...      % Corticospinal tract L
+                '9: ML_R', ...       % Medial lemniscus R
+                '10: ML_L', ...      % Medial lemniscus L
+                '11: ICP_R', ...     % Inferior cerebellar peduncle R
+                '12: ICP_L', ...     % Inferior cerebellar peduncle L
+                '13: SCP_R', ...     % Superior cerebellar peduncle R
+                '14: SCP_L', ...     % Superior cerebellar peduncle L
+                '15: CP_R', ...      % Cerebral peduncle R
+                '16: CP_L', ...      % Cerebral peduncle L
+                '17: ALIC_R', ...    % Anterior limb of internal capsule R
+                '18: ALIC_L', ...    % Anterior limb of internal capsule L
+                '19: PLIC_R', ...    % Posterior limb of internal capsule R
+                '20: PLIC_L', ...    % Posterior limb of internal capsule L
+                '21: RLIC_R', ...    % Retrolenticular part of internal capsule R
+                '22: RLIC_L', ...    % Retrolenticular part of internal capsule L
+                '23: ACR_R', ...     % Anterior corona radiata R
+                '24: ACR_L', ...     % Anterior corona radiata L
+                '25: SCR_R', ...     % Superior corona radiata R
+                '26: SCR_L', ...     % Superior corona radiata L
+                '27: PCR_R', ...     % Posterior corona radiata R
+                '28: PCR_L', ...     % Posterior corona radiata L
+                '29: PTR_R', ...     % Posterior thalamic radiation (include optic radiation) R
+                '30: PTR_L', ...     % Posterior thalamic radiation (include optic radiation) L
+                '31: SS_R', ...      % Sagittal stratum (incl ILF and IFOF) R
+                '32: SS_L', ...      % Sagittal stratum (incl ILF and IFOF) L
+                '33: EC_R', ...      % External capsule R
+                '34: EC_L', ...      % External capsule L
+                '35: CGC_R', ...     % Cingulum (cingulate gyrus) R
+                '36: CGC_L', ...     % Cingulum (cingulate gyrus) L
+                '37: CGH_R', ...     % Cingulum (hippocampus) R
+                '38: CGH_L', ...     % Cingulum (hippocampus) L
+                '39: FST_R', ...     % Fornix (crus) / Stria terminalis R
+                '40: FST_L', ...     % Fornix (crus) / Stria terminalis L
+                '41: SLF_R', ...     % Superior longitudinal fasciculus R
+                '42: SLF_L', ...     % Superior longitudinal fasciculus L
+                '43: SFOF_R', ...    % Superior fronto-occipital fasciculus R
+                '44: SFOF_L', ...    % Superior fronto-occipital fasciculus L
+                '45: UF_R', ...      % Uncinate fasciculus R
+                '46: UF_L', ...      % Uncinate fasciculus L
+                '47: TAP_R', ...     % Tapetum R
+                '48: TAP_L' ...      % Tapetum L
+                };
+
+
+
+
+            a = @(x) strsplit(x, ' ');
+            b = @(x,i) x{i};
+            c = @(x) str2num(x(1:(end-1)));
+
+            labels = cellfun(@(x) b(a(x),2), txt, 'UniformOutput', false);
+            ids = cellfun(@(x) c(b(a(x),1)), txt, 'UniformOutput', false);
+        end
+    end
+end
diff --git a/examples/my_nodes/dp_node_segm_xtractseg.m b/examples/my_nodes/dp_node_segm_xtractseg.m
new file mode 100644
index 0000000..5e071eb
--- /dev/null
+++ b/examples/my_nodes/dp_node_segm_xtractseg.m
@@ -0,0 +1,165 @@
+classdef dp_node_segm_xtractseg < dp_node_segm
+
+    % inputs
+    % dmri_fn
+    % bval_fn
+    % bvec_fn
+    % mask_fn
+
+    % to do: flow management, e.g. deleting files and more
+
+    methods
+
+        function obj = dp_node_segm_xtractseg()
+            obj.output_test = {'labels_fn'};
+        end
+
+        function output = i2o(obj, input)
+
+            output.op = input.op;
+
+            f = @(x) fullfile(output.op, x);
+            % output.mask_fn = f('nodif_brain_mask.nii.gz');
+            % output.dmri_fn = f('dmri.nii.gz');
+            % output.bval_fn = f('x.bvals');
+            % output.bvec_fn = f('x.bvecs');
+
+            output.mask_fn = f('nodif_brain_mask_MNI.nii.gz');
+            output.dmri_fn = f('Diffusion_MNI.nii.gz');
+            output.fa_fn = f('FA_MNI.nii.gz');
+            output.bval_fn = f('x.bvals');
+            output.bvec_fn = f('x.bvecs');
+
+
+
+            output.labels_fn = f('bundles.nii.gz');
+
+            output.tmp.bp = msf_tmp_path();
+            output.tmp.do_delete = 1;
+
+        end
+
+        function output = execute(obj, input, output)
+
+            % prepare
+            msf_mkdir(output.op);
+
+            % put brain mask and dmri into the right place
+            copyfile(input.dmri_fn, fullfile(output.op, 'dmri.nii.gz'));
+            copyfile(input.bval_fn, output.bval_fn);
+            copyfile(input.bvec_fn, output.bvec_fn);
+
+            % Preliminary implementation
+            if (~isempty(input.mask_fn))
+                copyfile(input.mask_fn, output.mask_fn);
+            end
+
+
+            % convert relative to absolute path
+            [~,tmp] = fileattrib(output.op);
+            op = tmp.Name;
+
+            % cmd = sprintf(...
+            %     'docker run -v "%s":/data -t "%s" TractSeg -i "%s" -o /data %s --bvals "%s" --bvecs "%s" %s', ...
+            %     op, ...
+            %     'wasserth/tractseg_container:master', ... % docker name
+            %     'data/dmri.nii.gz', ...
+            %     '--raw_diffusion_input', ... % options
+            %     'data/x.bvals', ...
+            %     'data/x.bvecs', ...
+            %     '--preprocess'); % register to mni space
+
+            cmd = sprintf(...
+                'docker run -v "%s":/data -t "%s" TractSeg -i "%s" -o /data %s --bvals "%s" --bvecs "%s" %s', ...
+                op, ...
+                'wasserth/tractseg_container:master', ...
+                'data/dmri.nii.gz', ...
+                '--raw_diffusion_input', ...
+                'data/x.bvals', ...
+                'data/x.bvecs', ...
+                '--preprocess --tract_definition xtract');
+
+
+            obj.syscmd(cmd);
+
+
+            % wrap up the files in a 4D volume
+            h = mdm_nii_read_header(output.dmri_fn);
+            labels = obj.segm_labels();
+            R = zeros(h.dim(2), h.dim(3), h.dim(4), numel(labels));
+            for c = 1:numel(labels)
+                tmp = mdm_nii_read(fullfile(output.op, ...
+                    'bundle_segmentations_MNI', ...
+                    cat(2, labels{c}, '.nii.gz')));
+                R(:,:,:,c) = tmp;
+            end
+            mdm_nii_write(uint8(R), output.labels_fn, h);
+
+        end
+
+    end
+
+    methods (Hidden)
+
+        function [labels, ids] = segm_info(obj)
+
+            % dp_node_segm_xtractseg.segm_info()  — XTRACT bundle names
+
+            txt = { ...
+                '0: ac', ...        % Anterior commissure
+                '1: af_l', ...      % Arcuate fasciculus (left)
+                '2: af_r', ...      % Arcuate fasciculus (right)
+                '3: ar_l', ...      % Acoustic radiation (left)
+                '4: ar_r', ...      % Acoustic radiation (right)
+                '5: atr_l', ...     % Anterior thalamic radiation (left)
+                '6: atr_r', ...     % Anterior thalamic radiation (right)
+                '7: cbd_l', ...     % Cingulum bundle – dorsal (left)
+                '8: cbd_r', ...     % Cingulum bundle – dorsal (right)
+                '9: cbp_l', ...     % Cingulum bundle – perigenual (left)
+                '10: cbp_r', ...    % Cingulum bundle – perigenual (right)
+                '11: cbt_l', ...    % Cingulum bundle – temporal (left)
+                '12: cbt_r', ...    % Cingulum bundle – temporal (right)
+                '13: cst_l', ...    % Corticospinal tract (left)
+                '14: cst_r', ...    % Corticospinal tract (right)
+                '15: fa_l', ...     % Frontal aslant tract (left)
+                '16: fa_r', ...     % Frontal aslant tract (right)
+                '17: fma', ...      % Forceps major (splenium of corpus callosum)
+                '18: fmi', ...      % Forceps minor (genu of corpus callosum)
+                '19: fx_l', ...     % Fornix (left)
+                '20: fx_r', ...     % Fornix (right)
+                '21: ifo_l', ...    % Inferior fronto-occipital fasciculus (left)
+                '22: ifo_r', ...    % Inferior fronto-occipital fasciculus (right)
+                '23: ilf_l', ...    % Inferior longitudinal fasciculus (left)
+                '24: ilf_r', ...    % Inferior longitudinal fasciculus (right)
+                '25: mcp', ...      % Middle cerebellar peduncle
+                '26: mdlf_l', ...   % Middle longitudinal fasciculus (left)
+                '27: mdlf_r', ...   % Middle longitudinal fasciculus (right)
+                '28: or_l', ...     % Optic radiation (left)
+                '29: or_r', ...     % Optic radiation (right)
+                '30: slf1_l', ...   % Superior longitudinal fasciculus I (left)
+                '31: slf1_r', ...   % Superior longitudinal fasciculus I (right)
+                '32: slf2_l', ...   % Superior longitudinal fasciculus II (left)
+                '33: slf2_r', ...   % Superior longitudinal fasciculus II (right)
+                '34: slf3_l', ...   % Superior longitudinal fasciculus III (left)
+                '35: slf3_r', ...   % Superior longitudinal fasciculus III (right)
+                '36: str_l', ...    % Superior thalamic radiation (left)
+                '37: str_r', ...    % Superior thalamic radiation (right)
+                '38: uf_l', ...     % Uncinate fasciculus (left)
+                '39: uf_r', ...     % Uncinate fasciculus (right)
+                '40: vof_l', ...    % Vertical occipital fasciculus (left)
+                '41: vof_r' ...     % Vertical occipital fasciculus (right)
+                };
+
+
+            a = @(x) strsplit(x, ' ');
+            b = @(x,i) x{i};
+            c = @(x) 1 + str2num(x(1:(end-1)));
+
+            labels = cellfun(@(x) b(a(x),2), txt, 'UniformOutput', false);
+            ids = cellfun(@(x) c(b(a(x),1)), txt, 'UniformOutput', false);
+
+
+        end
+    end
+end
+
diff --git a/examples/my_nodes/my_mask_trim.m b/examples/my_nodes/my_mask_trim.m
new file mode 100644
index 0000000..1b331b0
--- /dev/null
+++ b/examples/my_nodes/my_mask_trim.m
@@ -0,0 +1,53 @@
+classdef my_mask_trim < dp_node
+
+    methods
+
+        function output = i2o(obj, input)
+
+            output = input;
+            output.mask_fn = dp.new_fn(input.op, input.mask_fn, '_trim');
+
+
+            output.fa_fn = dp.new_fn(input.op, input.fa_fn, '_trim');
+            output.md_fn = dp.new_fn(input.op, input.md_fn, '_trim');
+            output.rd_fn = dp.new_fn(input.op, input.rd_fn, '_trim');
+            output.ad_fn = dp.new_fn(input.op, input.ad_fn, '_trim');
+            output.s0_fn = dp.new_fn(input.op, input.s0_fn, '_trim');
+            
+
+        end
+
+        function output = execute(obj, input, output)
+
+            % Load, trim, save the mask
+            [M,h] = mdm_nii_read(input.mask_fn);
+
+            f = @(x,a,b) mio_smooth_4d(double(x), a) > b;
+            M = f(f(f(M, 0.7, 0.3), 0.75, 0.8), 0.7, 0.85);
+            M = f(M, 0.7, 0.8);
+
+            mdm_nii_write(double(M), output.mask_fn, h);            
+
+            % Trim parameter maps
+
+            s = {'fa_fn', 'md_fn', 'ad_fn', 'rd_fn', 's0_fn'};
+
+            for c = 1:numel(s)
+
+                [I,h] = mdm_nii_read(input.(s{c}));
+                IM = medfilt3(I);
+
+                if (c == 1)
+                    ind = mio_smooth_4d(double(I - IM), 0.5) > 0.15;
+                end
+
+                I(ind) = IM(ind);
+
+                mdm_nii_write(I.*M, output.(s{c}), h);
+            end
+        end
+
+
+    end
+
+end
\ No newline at end of file
diff --git a/examples/my_nodes/my_primary_node.m b/examples/my_nodes/my_primary_node.m
new file mode 100644
index 0000000..cae0f5e
--- /dev/null
+++ b/examples/my_nodes/my_primary_node.m
@@ -0,0 +1,44 @@
+classdef my_primary_node < dp_node_primary
+
+    properties
+        bp;
+        pattern = '*';
+    end
+
+    methods
+
+        function obj = my_primary_node(bp, pattern)
+            obj.bp = bp;
+            obj.pattern = pattern;
+        end
+
+        function outputs = get_iterable(obj)
+
+            d = dir(fullfile(obj.bp, obj.pattern));
+
+            outputs = {};
+
+            % find folders
+            for c = 1:numel(d)
+
+                if (d(c).name(1) == '.')
+                    continue;
+                end
+
+                d2 = dir(fullfile(obj.bp, d(c).name, '2*'));
+
+                if (numel(d2) > 1)
+                    1;
+                end
+
+                for c2 = 1:numel(d2)
+
+                    output.bp = obj.bp;
+                    output.id = fullfile(d(c).name, d2(c2).name);
+
+                    outputs{end+1} = output; 
+                end
+            end
+        end
+    end
+end
\ No newline at end of file
diff --git a/examples/my_nodes/split_rois.m b/examples/my_nodes/split_rois.m
new file mode 100644
index 0000000..48fe6ab
--- /dev/null
+++ b/examples/my_nodes/split_rois.m
@@ -0,0 +1,95 @@
+classdef split_rois < dp_node_roi_from_label
+
+    properties
+        roi_list
+    end
+
+    methods
+
+
+        function obj = split_rois(name, segm_node, roi_list)
+            obj = obj@dp_node_roi_from_label(name,segm_node);
+            obj.roi_list = roi_list;
+
+            slab_suffix = {'_I','_II','_III'};
+
+            % find ROI ids from the segm node's vocabulary
+            base_ids = [];
+            for k = 1:numel(roi_list)
+                hit = find(strcmpi(obj.roi_names, roi_list{k}), 1);
+                if ~isempty(hit)
+                    % extract numeric value if roi_ids is a cell array
+                    if iscell(obj.roi_ids)
+                        base_ids(end+1) = obj.roi_ids{hit};
+                    else
+                        base_ids(end+1) = obj.roi_ids(hit);
+                    end
+                else
+                    warning('split_rois: tract %s not found', roi_list{k});
+                end
+            end
+
+            % assign three new IDs per base tract
+            next_id = max([obj.roi_ids{:}]) + 1;   % unwrap cells if needed
+
+
+            for i = 1:numel(base_ids)
+                base_name = obj.roi_names{i};   % or your roi_list{i}
+                for s = 1:3
+                    obj.roi_ids{end+1}   = next_id;
+                    obj.roi_names{end+1} = [roi_list{i} slab_suffix{s}];
+                    next_id = next_id + 1;
+                end
+            end
+
+        end
+
+        function [R,h] = roi_get_volume(obj, output, ~, c_roi)
+
+            rid = obj.roi_ids;
+            if iscell(rid), ridv = [rid{:}]; else, ridv = rid; end
+            ix = find(ridv == c_roi, 1);
+
+            name = obj.roi_names{ix};
+
+            % --- is this one of our synthetic slabs? (… I / II / III) ---
+            if endsWith(name,'_I') || endsWith(name,'_II') || endsWith(name,'_III')
+
+                % which slab is requested?
+                if     endsWith(name,'_I'),   slab = 1;
+                elseif endsWith(name,'_II'),  slab = 2;
+                else                          slab = 3;
+                end
+
+                % base tract name without the suffix
+                base_name = regexprep(name, '(_I|_II|_III)$', '');
+
+                % look up the base tract's ROI id by exact (case-insensitive) name
+                ix_base = find(strcmpi(obj.roi_names, base_name), 1);
+
+                if iscell(rid), base_id = rid{ix_base}; else, base_id = rid(ix_base); end
+
+                % fetch the base mask via the parent implementation
+                [R,h] = roi_get_volume@dp_node_roi_from_label(obj, output, 1, base_id);
+
+                switch slab
+                    case 1
+                        R(:, :, 44:end) = 0;
+                    case 2
+                        R(:, :, 1:44) = 0;
+                        R(:, :, 66:end) = 0;
+                    case 3
+                        R(:, :, 1:66) = 0;
+                end
+
+            else
+                [R,h] = roi_get_volume@dp_node_roi_from_label(obj, output, 1, c_roi);
+            end
+
+
+        end
+
+    end
+
+
+end
\ No newline at end of file
diff --git a/examples/step2_pipe_tractseg_xtract_juelich.m b/examples/step2_pipe_tractseg_xtract_juelich.m
new file mode 100644
index 0000000..d9aaeca
--- /dev/null
+++ b/examples/step2_pipe_tractseg_xtract_juelich.m
@@ -0,0 +1,211 @@
+% Root to NIfTI data
+base_path = '/path/to/your/data/nii';
+
+% Subjects → add dmri/xps/op
+id1 = my_primary_node(base_path, '*');
+% id1 = dp_node_io_filter_by_number(1).connect(id1);  % keep only first n subjects
+
+id2 = dp_node_io_append({...
+  {'dmri_fn', @(x) fullfile(x.bp, x.id, 'DWI', 'DTI_dn_mc.nii.gz')}, ...
+  {'xps_fn',  @(x) fullfile(x.bp, x.id, 'DWI', 'DTI_dn_mc_xps.mat')}, ...
+  {'op',      @(x) fullfile(x.bp, x.id, 'tractseg')}}).connect(id1);
+
+% b0 → mask
+m0 = dp_node_dmri_subsample_b0().connect(id2);
+m1 = dp_node_segm_hd_bet().connect(m0, 'm1');
+
+% xps→bval/bvec, gradcheck
+t0 = dp_node_dmri_io_xps_to_bval_bvec().connect(id2, 't0');
+t1 = dp_node_mrtrix_dwigradcheck().connect(t0);
+
+% merge mask + grads, alias mask
+t2 = dp_node_io_merge({t1, m1}); t2.do_prefix = 0;
+t3 = dp_node_io('mask_fn', 'm1_mask_fn').connect(t2);
+
+% TractSeg (MNI outputs: Diffusion_MNI, FA_MNI, nodif_brain_mask_MNI, bundles.nii.gz)
+t4 = dp_node_segm_tractseg().connect(t3);
+
+% XTRACT definitions in a separate output folder
+xt_op = dp_node_io('op', @(x) fullfile(x.bp, x.id, 'tractseg_xtract')).connect(t3);
+xt1 = dp_node_segm_xtractseg().connect(xt_op);
+
+% xt1.run('report');
+% xt1.run('execute');
+% xt1.run('execute', struct('do_overwrite',1));
+
+% DTI (LLS) on TractSeg’s MNI-space diffusion data (b_delta = 1) into subfolder <op>/dti_lls
+dti1 = dp_node_dmri_xps_from_bval_bvec(1).connect(t4);
+dti2 = dp_node_io('op', @(x) fullfile(x.bp, x.id, 'tractseg', 'dti_lls')).connect(dti1);
+dti3 = dp_node_dmri_dti().connect(dti2);
+
+% Refine brain mask and trim FA/MD/AD/RD/S0 maps using smoothing + median filtering
+dti_trim = my_mask_trim().connect(dti3);
+
+
+% Register AD/RD output filenames for dti_lls (not used downstream in this script)
+% dti3_append = dp_node_io_append({
+%   {'ad_fn', @(x) fullfile(x.op, 'dti_lls_ad.nii.gz')}, ...
+%   {'rd_fn', @(x) fullfile(x.op, 'dti_lls_rd.nii.gz')}
+% }).connect(dti3);
+
+
+% Merge trimmed DTI maps with TractSeg outputs (keep original field names)
+mdt = dp_node_io_merge({dti_trim, t4}); mdt.do_prefix = 0;
+
+% Expose only labels + FA/MD/AD/RD for TractSeg ROI stats (used by t5, t5_split)
+troi = dp_node_io_rename({
+  {'labels_fn','dp_node_segm_tractseg_labels_fn'}, ...
+  {'fa_fn','fa_fn'}, ...
+  {'md_fn','md_fn'}, ...
+  {'ad_fn','ad_fn'}, ...
+  {'rd_fn','rd_fn'} ...
+}).connect(mdt);
+
+% Merge trimmed DTI maps with XTRACT-based TractSeg outputs (keep original field names)
+mdtx = dp_node_io_merge({dti_trim, xt1}); mdtx.do_prefix = 0;
+
+% Expose only labels + FA/MD/AD/RD for XTRACT ROI stats (used by xt2, xt1_split)
+xtroi = dp_node_io_rename({
+  {'labels_fn','dp_node_segm_xtractseg_labels_fn'}, ...
+  {'fa_fn','fa_fn'}, ...
+  {'md_fn','md_fn'}, ...
+  {'ad_fn','ad_fn'}, ...
+  {'rd_fn','rd_fn'} ...
+}).connect(mdtx);
+
+
+% TractSeg bundle ROI stats → CSV (mean + voxel count)
+t5 = dp_node_roi_from_label('tractseg', t4).connect(troi);
+t_csv = dp_node_csv('tractseg_export','../report/csv', {'mean','n'}).connect(t5);
+
+% TractSeg: split selected bundles into 3 z-slabs (I/II/III) and export ROI stats
+ts_split_list = { ...
+    'CST_left','CST_right', ...
+    'SCP_left','SCP_right', ...
+    'ICP_left','ICP_right', ...
+    'CC_1','CC_2','CC_3','CC_4','CC_5','CC_6','CC_7', ...
+    'MCP', ...
+    'IFO_left','IFO_right', ...
+    'ILF_left','ILF_right', ...
+    'CG_left','CG_right', ...
+    'UF_left','UF_right'};
+
+% TractSeg: split selected bundles into 3 z-slabs (I/II/III) and export ROI stats → CSV (mean + voxel count)
+t5_split  = split_rois('tractseg_slab', t4, ts_split_list).connect(troi);
+t6_split  = dp_node_csv('tractseg_export_with_slabs','../report/csv', {'mean','n'}).connect(t5_split);
+
+% XTRACT-based TractSeg bundle ROI stats → CSV (mean + voxel count)
+xt2 = dp_node_roi_from_label('tractseg_xtract', xt1).connect(xtroi);
+xt3 = dp_node_csv('tractseg_xtract_export','../report/csv', {'mean','n'}).connect(xt2);
+
+% XTRACT: split CST (left/right) into 3 z-slabs and export ROI stats → CSV (mean + voxel count)
+xt1_split = split_rois('tractseg_xtract_slab', xt1, {'cst_l','cst_r'}).connect(xtroi);
+xt_csv = dp_node_csv('tractseg_xtract_export_with_slabs','../report/csv', {'mean','n'}).connect(xt1_split);
+
+
+% ------------- JHU ROIs using ANTs (quick registration) ----------------------
+%
+% TractSeg places dMRI in a semi-MNI rigid space (t4.*_MNI).
+% DTI maps (dti3) are computed on that same grid.
+% To align the JHU atlas with this subject grid:
+%   (A) Register subject FA (semi-MNI) → JHU FA template using ANTs quick mode
+%   (B) Apply the inverse warp to bring JHU labels → subject trimmed FA grid
+%   (C) Crop warped labels by the subject mask
+%
+% Result: JHU atlas labels and trimmed DTI maps are voxel-aligned on the semi-MNI (TractSeg) grid.
+
+
+% JHU atlas paths
+jhu_fa_fn     = '/usr/local/fsl/data/atlases/JHU/JHU-ICBM-FA-1mm.nii.gz';
+jhu_labels_fn = '/usr/local/fsl/data/atlases/JHU/JHU-ICBM-labels-1mm.nii.gz';
+
+% ----------- ANTs quick registration branch --------
+amni0 = dp_node_io('op', @(x) fullfile(x.bp, x.id, 'tractseg_jhu_ants_quick')).connect(dti_trim);
+% amni0.run('report');
+
+% ANTs: subject trimmed FA (moving) -> JHU FA (fixed) 
+amni1 = dp_node_io('nii_fn', 'fa_fn').connect(amni0); % moving: trimmed FA
+% amni1.run('report');
+
+amni2 = dp_node_io('target_fn', @(x) jhu_fa_fn).connect(amni1); % fixed: JHU FA
+% amni2.run('report');
+
+amni3 = dp_node_ants_reg_quick().connect(amni2);   
+% amni3.run('report');
+
+% Attach trimmed FA grid as reference for antsApply
+amni4 = dp_node_io_merge({amni3, dti_trim}); 
+amni4.do_prefix = 0;
+
+% Use trimmed FA as ref_fn
+amni5 = dp_node_io_append({{'ref_fn','my_mask_trim_fa_fn'}}).connect(amni4);
+
+% Warp full JHU labels to subject trimmed FA space (using inverse chain)
+amni6 = dp_node_io('nii_fn', @(x) jhu_labels_fn).connect(amni5);
+amni7 = dp_node_ants_apply(' -n NearestNeighbor').connect(amni6);
+% amni7.nii_fn: JHU labels (quick) on trimmed FA grid
+
+% Crop warped labels by subject mask on same grid 
+amni_merge1  = dp_node_io_merge({amni7, dti_trim}); amni_merge1.do_prefix = 0;
+amni_merge2 = dp_node_io_append({{'mask_fn','my_mask_trim_fa_fn'}}).connect(amni_merge1);
+jhu_crop = dp_node_atlas_crop_to_mask().connect(amni_merge2);
+
+
+% QC
+if (0)
+    qc1 = dp_node_io_merge({dti_trim, amni3, jhu_crop}); qc1.do_prefix = 0;
+
+    % subject trimmed FA + labels warped back from MNI to subject space
+    if (0)
+        qc2 = dp_node_io_rename({
+            {'fa_fn',     'fa_fn'}, ...
+            {'labels_fn', 'dp_node_atlas_crop_to_mask_nii_fn'}, ...
+            {'mask_fn',   'mask_fn'}
+            }).connect(qc1);
+    end
+
+    % subject FA warped into JHU (MNI) space + labels in MNI
+    if (0)
+        qc2 = dp_node_io_rename({
+            {'fa_fn',     'dp_node_ants_reg_quick_warped_fn'}, ...
+            {'labels_fn', @(x) jhu_labels_fn}
+            }).connect(qc1);
+    end
+
+    % JHU template FA + JHU atlas labels (reference anatomy in MNI space)
+    if (0)
+        qc2 = dp_node_io_rename({
+            {'fa_fn',     @(x) jhu_fa_fn}, ...
+            {'labels_fn', @(x) jhu_labels_fn}
+            }).connect(qc1);
+    end
+
+    % qc2.run('mgui');
+
+    qc3 = dp_node_roi_from_label('tmp', dp_node_segm_jhu()).connect(qc2);
+    qc3.run('mgui');
+end
+
+
+% ****** Merge for ROI stats ******
+
+mda = dp_node_io_merge({dti_trim, jhu_crop});
+mda.do_prefix = 0;
+
+aroi = dp_node_io_rename({
+  {'labels_fn','dp_node_atlas_crop_to_mask_nii_fn'}, ...
+  {'fa_fn','fa_fn'}, ...
+  {'md_fn','md_fn'}, ...
+  {'ad_fn','ad_fn'}, ...
+  {'rd_fn','rd_fn'}
+}).connect(mda);
+
+
+jhu_LUT  = dp_node_segm_jhu();  %LUT
+jroi    = dp_node_roi_from_label('jhu_ants_quick', jhu_LUT).connect(aroi);
+j_cvs = dp_node_csv('jhu_ants_export','../report/csv', {'mean','n'}).connect(jroi);
+
+
+
+
+