A single computational objective can produce specialization of streams in visual cortex.
11 blocks of code here have a permanent identifier (oscr:b:…), listed with the matches below: citable for good, with a page of its own. Select a match and press Cite to copy it. What a block identifier is.
The 16 matches · 1 of them tie a paragraph to a whole file without given lines: a weak match, whose lines are not tinted
- [1] § Results › Stream-hypothesized task models do not yield stream-specific spatial or functional correspondence ↔ notebooks/demo_plot_combined_source_data.ipynb, lines 333 to 399 · score 0.88 · MB v3 categorization, ConvNeXT, SlowFast, MB v2, MB v1, ResNet · Report a mismatch
- [2] § Results › Computational framework for stream-organization hypotheses ↔ notebooks/demo_plot_combined_source_data.ipynb, lines 333 to 399 · score 0.87 · DepthAnythingv2, ConvNeXT, DEKR Pose, MB v3, MB v2, MB v1 · Report a mismatch
- [3] § Methods › Representational overlap across MB v1 task models ↔ scripts/cka_mb18_overlap.py, lines 1 to 45 · score 0.87 · backbone.feature_provider.feature_provid, slow.layer4.1, linear CKA, overlap, detection, stimuli · block
oscr:b:CEK6RTA5XWCZS-J· Report a mismatch - [4] § Methods › Representational overlap across MB v1 task models ↔ scripts/fitting_one_to_one_unit2voxel.py, lines 30 to 112 · score 0.77 · backbone.feature_provider.feature_provid, slow.layer4.1, detection, v1, layers, models · block
oscr:b:XK72NT8ZRBW96-9· Report a mismatch - [5] § Methods › Model unit selectivity and receptive field properties ↔ spacetorch/datasets/floc.py, lines 79 to 200 · score 0.69 · category selective, corridors, houses, instruments, limbs, adult · block
oscr:b:G2MXPJC6ANCRZ-P· Report a mismatch - [6] § Results › Stream-hypothesized task models do not yield stream-specific spatial or functional correspondence ↔ matlab/F04_A.m, lines 86 to 103 · score 0.67 · ConvNeXT, SlowFast, ResNet, CNN, SSD, Faster · block
oscr:b:RF89V6745WRVK-9· Report a mismatch - [7] § Methods › Training: Neural network architectures and training tasks › Multiple behaviors models ↔ spacestream/utils/get_utils.py, lines 206 to 287 · score 0.66 · SlowFast model, detection model, Kinetics, vision, Faster, backbone · block
oscr:b:VB50QMEK8A51Z-6· Report a mismatch - [8] § Methods › Testing: Evaluating theories › Linear regression ↔ spacestream/utils/mapping_utils.py, lines 17 to 84 · score 0.62 · linear regression, voxel responses, ridge, splits, mapping, layer · block
oscr:b:9RXZEFKB9D5DR-X· Report a mismatch - [9] § Results › Stream-hypothesized task models do not yield stream-specific spatial or functional correspondence ↔ notebooks/F04_A_processing.ipynb, lines 129 to 147 · score 0.59 · v3 depth, v3 pose, v3 categorization, MB, dorsal, lateral · Report a mismatch
- [10] § Methods › Training: Neural network architectures and training tasks › Spatial constraints models › Loss functions ↔ spacestream/models/slow_fast.py, lines 10 to 76 · score 0.59 · cross entropy loss, loss function, network, layer, model · block
oscr:b:JAACS7ZA66KWJ-K· Report a mismatch - [11] § Methods › Testing: Evaluating theories › Brain data › Category selectivity by stream in human cortex ↔ matlab/transform_selectivity_to_fsaverage.m, the whole file · a weak match · score 0.59 · fsaverage space, contrast maps, selectivity, bodies, fLoc, NSD · block
oscr:b:PVZT8GSKN44CZ-I· Report a mismatch - [12] § Results › Stream-hypothesized task models do not yield stream-specific spatial or functional correspondence ↔ notebooks/F03_B_processing.ipynb, lines 155 to 164 · score 0.55 · MB v2, MB v1, multiple behavior, collapsed, hypothesized, models · Report a mismatch
- [13] § Methods › Training: Neural network architectures and training tasks › Multiple behaviors models ↔ spacestream/utils/get_utils.py, lines 206 to 287 · score 0.55 · OpenAI, encoder, vision, Pretraining, ConvNeXT, DETR · block
oscr:b:VB50QMEK8A51Z-6· Report a mismatch - [14] § Methods › Training: Neural network architectures and training tasks › Spatial constraints models › Loss functions ↔ spacetorch/losses/cross_entropy_spatial_correlation_loss.py, lines 11 to 58 · score 0.54 · cross entropy loss, batch, layer · block
oscr:b:PHSXR1GRY51YB-K· Report a mismatch - [15] § Methods › Training: Neural network architectures and training tasks › Spatial constraints models › Initialization of model unit position ↔ spacetorch/datasets/__init__.py, lines 1 to 35 · score 0.54 · retinal waves, sine grating · block
oscr:b:B1ZMF48S1SSTX-9· Report a mismatch - [16] § Methods › Testing: Evaluating theories › Linear regression ↔ spacestream/analyses/base_fitter.py, lines 102 to 209 · score 0.52 · lower dimensional space, regression, splits, layer, train, stream · block
oscr:b:VZH4GT9BX0YZ0-H· Report a mismatch
Paper
Loaded from Europe PMC by your browser; OSCR does not store it: doi.org · Europe PMC
The paper is loaded when this pane is shown.
The authors' code
Jupyter notebook · 991 lines · 41 KB · Apache-2.0 · 2 matches
- # %% [markdown]
- # # Demo: Plot Combined Manuscript Source Data
- #
- # This notebook reads only `data/spacestream_manuscript_source_data.xlsx`. Each worksheet is one manuscript figure panel; panels with multiple source tables are parsed from stacked sections within that worksheet.
- # %%
- from pathlib import Path
- import os
- import zipfile
- import xml.etree.ElementTree as ET
- import matplotlib.pyplot as plt
- from matplotlib.lines import Line2D
- import numpy as np
- import pandas as pd
- import seaborn as sns
- ROOT = Path.cwd()
- if ROOT.name != "manuscript_figure_source_data":
- ROOT = Path("/oak/stanford/groups/kalanit/biac2/kgs/projects/Dawn/manuscript_figure_source_data")
- WORKBOOK = ROOT / "data" / "spacestream_manuscript_source_data.xlsx"
- OUT = ROOT / "outputs" / "combined_demo_plots"
- OUT.mkdir(parents=True, exist_ok=True)
- os.environ.setdefault("MPLCONFIGDIR", str(ROOT / ".matplotlib"))
- os.environ.setdefault("XDG_CACHE_HOME", str(ROOT / ".cache"))
- (ROOT / ".matplotlib").mkdir(exist_ok=True)
- (ROOT / ".cache").mkdir(exist_ok=True)
- plt.rcParams.update({
- "figure.dpi": 120,
- "savefig.dpi": 200,
- "axes.spines.top": False,
- "axes.spines.right": False,
- "font.size": 10,
- "font.family": "sans-serif",
- "font.sans-serif": ["Helvetica", "Arial", "DejaVu Sans"],
- "pdf.use14corefonts": True,
- "ps.useafm": True,
- "svg.fonttype": "none",
- "pdf.fonttype": 42,
- "ps.fonttype": 42,
- })
- sns.set_theme(style="ticks", rc={
- "font.family": "sans-serif",
- "font.sans-serif": ["Helvetica", "Arial", "DejaVu Sans"],
- })
- CORE_ROI_NAMES = ["Ventral", "Lateral", "Dorsal"]
- ROI_ORDER = ["Dorsal", "Lateral", "Ventral"]
- ROI_COLORS = ["#377E2C", "#1A1AAC", "#8C1A4C"]
- STREAM_PALETTE = {"Ventral": ROI_COLORS[2], "Lateral": ROI_COLORS[1], "Dorsal": ROI_COLORS[0]}
- STREAM_COLORS = {
- "Dorsal": "#006600",
- "Parietal": "#006600",
- "Lateral": "#4d7fff",
- "Ventral": "#990000",
- }
- STREAM_ORDER = ["Dorsal", "Lateral", "Ventral"]
- LEGACY_STREAM_ORDER = ["Parietal", "Lateral", "Ventral"]
- CONTRAST_ORDER = ["Places", "Bodies", "Faces"]
- TDANN_COLORS = {"simCLR": "#720298", "self_supervised": "#720298", "supervised": "#CB6D4A"}
- def savefig(fig, stem):
- png = OUT / f"{stem}.png"
- pdf = OUT / f"{stem}.pdf"
- fig.tight_layout()
- fig.savefig(png)
- fig.savefig(pdf)
- plt.close(fig)
- print(png.relative_to(ROOT))
- print(pdf.relative_to(ROOT))
- def despine(ax):
- ax.spines["top"].set_visible(False)
- ax.spines["right"].set_visible(False)
- def clean_legend_to_streams(ax, stream_order):
- handles, labels = ax.get_legend_handles_labels()
- stream_handles = {}
- for handle, label in zip(handles, labels):
- if label in stream_order and label not in stream_handles:
- stream_handles[label] = handle
- if stream_handles:
- ax.legend(stream_handles.values(), stream_handles.keys(), frameon=False, title="Stream")
- else:
- ax.legend([], [], frameon=False)
- # %% [markdown]
- # ## Load Combined Panel Workbook
- # %%
- NS_MAIN = "{http://schemas.openxmlformats.org/spreadsheetml/2006/main}"
- NS_REL = "{http://schemas.openxmlformats.org/package/2006/relationships}"
- NS_DOC_REL = "{http://schemas.openxmlformats.org/officeDocument/2006/relationships}"
- def _read_shared_strings(zf):
- if "xl/sharedStrings.xml" not in zf.namelist():
- return []
- root = ET.fromstring(zf.read("xl/sharedStrings.xml"))
- return ["".join(node.text or "" for node in item.findall(f".//{NS_MAIN}t")) for item in root.findall(f"{NS_MAIN}si")]
- def _read_sheet_rows(zf, sheet_path, shared_strings):
- root = ET.fromstring(zf.read(sheet_path))
- rows = []
- for row in root.findall(f".//{NS_MAIN}row"):
- values = []
- expected_col = 1
- for cell in row.findall(f"{NS_MAIN}c"):
- ref = cell.attrib.get("r", "A1")
- col_idx = 0
- for ch in "".join(ch for ch in ref if ch.isalpha()):
- col_idx = col_idx * 26 + (ord(ch.upper()) - 64)
- while expected_col < col_idx:
- values.append("")
- expected_col += 1
- cell_type = cell.attrib.get("t")
- if cell_type == "inlineStr":
- text_node = cell.find(f"{NS_MAIN}is/{NS_MAIN}t")
- values.append(text_node.text if text_node is not None and text_node.text is not None else "")
- elif cell_type == "s":
- value_node = cell.find(f"{NS_MAIN}v")
- idx = int(value_node.text) if value_node is not None and value_node.text is not None else -1
- values.append(shared_strings[idx] if 0 <= idx < len(shared_strings) else "")
- else:
- value_node = cell.find(f"{NS_MAIN}v")
- values.append(value_node.text if value_node is not None and value_node.text is not None else "")
- expected_col += 1
- while values and values[-1] == "":
- values.pop()
- rows.append(values)
- return rows
- def _numericize(df):
- out = df.copy()
- for col in out.columns:
- try:
- out[col] = pd.to_numeric(out[col])
- except (TypeError, ValueError):
- pass
- return out
- def load_source_workbook(path):
- with zipfile.ZipFile(path) as zf:
- shared_strings = _read_shared_strings(zf)
- workbook_xml = ET.fromstring(zf.read("xl/workbook.xml"))
- rels_xml = ET.fromstring(zf.read("xl/_rels/workbook.xml.rels"))
- rel_targets = {rel.attrib["Id"]: rel.attrib["Target"] for rel in rels_xml.findall(f"{NS_REL}Relationship")}
- panel_rows = {}
- for sheet in workbook_xml.findall(f".//{NS_MAIN}sheet"):
- name = sheet.attrib["name"]
- rid = sheet.attrib[f"{NS_DOC_REL}id"]
- sheet_path = "xl/" + rel_targets[rid].lstrip("/")
- panel_rows[name] = _read_sheet_rows(zf, sheet_path, shared_strings)
- return panel_rows
- def extract_panel_tables(panel_rows):
- panel_tables = {}
- tables = {}
- csvs = {}
- metadata_keys = {"figure_panel", "source_workbook", "source_table", "staging_csv", "rows"}
- for panel, rows in panel_rows.items():
- panel_tables[panel] = {}
- i = 0
- while i < len(rows):
- row = rows[i]
- if row and row[0] == "source_table":
- table_name = row[1]
- staging_csv = ""
- i += 1
- while i < len(rows):
- row = rows[i]
- if row and row[0] == "staging_csv":
- staging_csv = row[1] if len(row) > 1 else ""
- if row and row[0] not in metadata_keys:
- break
- i += 1
- if i >= len(rows):
- break
- header = rows[i]
- data_rows = []
- i += 1
- while i < len(rows):
- row = rows[i]
- if row and row[0] == "source_table":
- break
- if row:
- data_rows.append(row)
- i += 1
- width = len(header)
- padded = [r + [""] * max(width - len(r), 0) for r in data_rows]
- df = _numericize(pd.DataFrame([r[:width] for r in padded], columns=header))
- panel_tables[panel][table_name] = df
- tables[table_name] = df
- if staging_csv:
- csvs[Path(staging_csv).stem] = df
- continue
- i += 1
- return panel_tables, tables, csvs
- panel_rows = load_source_workbook(WORKBOOK)
- panel_tables, sheets, csvs = extract_panel_tables(panel_rows)
- print(f"Loaded {len(panel_tables)} panel sheets from {WORKBOOK}")
- print(f"Extracted {len(sheets)} source tables")
- for panel, table_map in panel_tables.items():
- print(panel, sorted(table_map))
- # %% [markdown]
- # ## Fig 1A: Mega Matrix And MDS
- # %%
- coords = csvs["fig1a_mds_coordinates_random_state0"].copy()
- mega_matrix = csvs["fig1a_mega_matrix"].to_numpy(dtype=float)
- roi_colors = {
- "Early": "#a6a6a6",
- "Midventral": "#f4bdd8",
- "Midlateral": "#ccdaff",
- "Midparietal": "#b3ffc6",
- "Ventral": "#DC267F",
- "Lateral": "#4d7fff",
- "Parietal": "#006600",
- }
- markers = {"lh": "v", "rh": "o"}
- fig, ax = plt.subplots(figsize=(10, 10))
- for (roi, hemi), group in coords.groupby(["roi_raw", "hemi"]):
- ax.scatter(
- group["mds_dim1"],
- group["mds_dim2"],
- s=35 + 20 * group["subject"].astype(int),
- c=roi_colors[roi],
- marker=markers[hemi],
- edgecolors="white",
- linewidths=0.5,
- alpha=0.9,
- label=f"{roi}, {hemi}",
- )
- ax.set_xlabel("MDS dimension 1")
- ax.set_ylabel("MDS dimension 2")
- ax.set_title("Fig 1A MDS from saved mega matrix")
- ax.spines[["top", "right"]].set_visible(False)
- savefig(fig, "fig1a_mds_source_style_demo")
- fig, ax = plt.subplots(figsize=(8, 8))
- im = ax.imshow(mega_matrix, cmap="magma", vmin=0, vmax=1)
- ax.set_title("Fig 1A source mega matrix")
- ax.set_xlabel("Matrix index")
- ax.set_ylabel("Matrix index")
- fig.colorbar(im, ax=ax, fraction=0.046, pad=0.04)
- savefig(fig, "fig1a_mega_matrix_demo")
- # %% [markdown]
- # ## Fig 1B: Human fLoc Selectivity
- # %%
- def plot_floc_bar_strip(subject_df, hemi_df, stem, title, hue_order=None):
- sns.set_theme(style="ticks", context="talk")
- hue_order = hue_order or ["Dorsal", "Lateral", "Ventral"]
- fig, ax = plt.subplots(figsize=(6, 6), facecolor="white")
- ax.set_facecolor("white")
- sns.barplot(
- data=subject_df,
- x="contrast",
- y="frac_t_above",
- hue="stream",
- order=CONTRAST_ORDER,
- hue_order=hue_order,
- palette=STREAM_PALETTE,
- ci="sd",
- capsize=0.08,
- errwidth=1.5,
- edgecolor="white",
- linewidth=4,
- ax=ax,
- )
- markers = {"lh": "o", "rh": "^"}
- for hemi, marker in markers.items():
- sns.stripplot(
- data=hemi_df[hemi_df["hemi"] == hemi],
- x="contrast",
- y="frac_t_above",
- hue="stream",
- order=CONTRAST_ORDER,
- hue_order=hue_order,
- dodge=True,
- jitter=0.15,
- marker=marker,
- size=4.5,
- edgecolor="white",
- linewidth=0.5,
- palette=STREAM_PALETTE,
- ax=ax,
- )
- clean_legend_to_streams(ax, hue_order)
- ax.spines["right"].set_visible(False)
- ax.spines["top"].set_visible(False)
- ax.set_xlabel("Contrast")
- ax.set_ylabel("Fraction t > 3")
- ax.set_title(title)
- ax.grid(False)
- ax.tick_params(colors="black", labelcolor="black")
- ax.spines["left"].set_color("black")
- ax.spines["bottom"].set_color("black")
- savefig(fig, stem)
- plot_floc_bar_strip(
- csvs["fig1b_human_floc_selectivity_subject_bars"],
- csvs["fig1b_human_floc_selectivity_hemi_dots"],
- "fig1b_human_floc_selectivity_source_style_demo",
- "Fig 1B human fLoc selectivity",
- hue_order=["Dorsal", "Lateral", "Ventral"],
- )
- # %% [markdown]
- # ## Figs 3B/3C/4A/4B: Model Correspondence Panels
- # %%
- FIG3_MODEL_ORDER = [
- "MB_RN50", "MB_RN50_v2", "MB_RN50_v3", "MB_RN18",
- "TDANN_Supervised_0.0", "TDANN_SimCLR_0.0", "TDANN_Supervised", "TDANN_SimCLR",
- ]
- FIG3_MODEL_LABELS = ["MB v1", "MB v2", "MB v3", "MB v1", "Cat", "SimCLR", "TDANN Cat", "TDANN SimCLR"]
- FIG3_IN_BAR_LABELS = ["50", "50", "50", "18", "18", "18", "18", "18"]
- FIG3_XVALS = [1, 2, 3, 4.5, 5.5, 6.5, 8, 9]
- FIG4_MODEL_ORDER = [
- "MB_RN50_detection", "MB_RN50_action", "MB_RN50_categorization",
- "MB_RN50_v2_detection", "MB_RN50_v2_clip", "MB_RN50_v2_categorization",
- "MB_RN50_v3_depth", "MB_RN50_v3_pose", "MB_RN50_v3_categorization",
- "MB_RN18_detection", "MB_RN18_action", "MB_RN18_categorization",
- ]
- FIG4_MODEL_LABELS = ["SSD", "SlowFast", "ResNet", "DETR", "CLIP", "ConvNeXT-T", "DepthAnythingv2", "DEKR-Pose", "ResNet", "Faster R-CNN", "SlowFast", "ResNet"]
- FIG4_IN_BAR_LABELS = ["50", "50", "50", "50", "50", "50", "50", "50", "50", "18", "18", "18"]
- FIG4_XVALS = [1, 2, 3, 5, 6, 7, 9, 10, 11, 13, 14, 15]
- MATLAB_BAR_COLORS_FIG3 = ["#808080", "#808080", "#808080", "#808080", "#cc3300", "#660091", "#cc3300", "#660091"]
- MATLAB_BAR_ALPHAS_FIG3 = [0.35, 0.35, 0.35, 0.5, 0.5, 0.5, 0.75, 1.0]
- MATLAB_BAR_COLORS_FIG4 = ["#80b380", "#8080b3", "#cc8066", "#80b380", "#6666e6", "#cc8066", "#66cc66", "#4d4de6", "#cc8066", "#80b380", "#8080b3", "#cc8066"]
- def plot_matlab_like_three_roi(points, noise, model_order, model_labels, in_bar_labels, xvals, colors, alphas, stem, title, ylabel):
- noise_summary = noise.groupby("ROI")["result"].agg(["mean", "std"])
- ymax = float((noise_summary["mean"] + noise_summary["std"]).max() * 1.1)
- fig, axes = plt.subplots(1, 3, figsize=(16, 6), sharey=True)
- for ax, roi in zip(axes, ROI_ORDER):
- roi_points = points[points["ROIS"] == roi]
- band = noise_summary.loc[roi]
- ax.axhspan(band["mean"] - band["std"], band["mean"] + band["std"], color="0.9", alpha=1.0, zorder=0)
- means = roi_points.groupby("model_type")["result"].mean()
- for i, model in enumerate(model_order):
- mean = means.loc[model] if model in means.index else np.nan
- ax.bar(xvals[i], mean, width=0.95, color=colors[i], alpha=alphas[i], edgecolor="none", zorder=2)
- if np.isfinite(mean):
- ax.text(xvals[i], ymax * 0.04, in_bar_labels[i], rotation=90, color="0.82", fontsize=13, ha="center", va="bottom")
- model_rows = roi_points[roi_points["model_type"] == model]
- lh = model_rows[model_rows["hemi"] == "lh"]["result"]
- rh = model_rows[model_rows["hemi"] == "rh"]["result"]
- edge = {"Dorsal": "#003300", "Lateral": "#000033", "Ventral": "#330018"}[roi]
- ax.scatter(np.full(len(lh), xvals[i]), lh, marker="^", facecolors="none", edgecolors=edge, linewidths=1.4, zorder=3)
- ax.scatter(np.full(len(rh), xvals[i]), rh, marker="o", facecolors="none", edgecolors=edge, linewidths=1.4, zorder=3)
- ax.set_xticks(xvals)
- ax.set_xticklabels(model_labels, rotation=90, fontsize=10)
- ax.set_xlim(0, max(xvals) + 1.5)
- ax.set_ylim(0, ymax)
- ax.set_title(roi, fontsize=18)
- ax.spines["right"].set_visible(False)
- ax.spines["top"].set_visible(False)
- if ax is axes[0]:
- ax.set_ylabel(ylabel, fontsize=14)
- else:
- ax.tick_params(labelleft=False)
- ax.spines["left"].set_visible(False)
- fig.suptitle(title, fontsize=18)
- savefig(fig, stem)
- plot_matlab_like_three_roi(csvs["fig3b_spatial_correspondence_points"], csvs["fig3b_spatial_correspondence_noise_ceiling"], FIG3_MODEL_ORDER, FIG3_MODEL_LABELS, FIG3_IN_BAR_LABELS, FIG3_XVALS, MATLAB_BAR_COLORS_FIG3, MATLAB_BAR_ALPHAS_FIG3, "fig3b_spatial_correspondence_source_style_demo", "Fig 3B spatial correspondence", "Distance similarity [r]")
- plot_matlab_like_three_roi(csvs["fig3c_functional_correspondence_points"], csvs["fig3c_functional_correspondence_noise_ceiling"], FIG3_MODEL_ORDER, FIG3_MODEL_LABELS, FIG3_IN_BAR_LABELS, FIG3_XVALS, MATLAB_BAR_COLORS_FIG3, MATLAB_BAR_ALPHAS_FIG3, "fig3c_functional_correspondence_source_style_demo", "Fig 3C functional correspondence", "Functional similarity [r]")
- plot_matlab_like_three_roi(csvs["fig4a_mb_task_stream_assignment_points"], csvs["fig4a_mb_task_stream_assignment_noise_ceiling"], FIG4_MODEL_ORDER, FIG4_MODEL_LABELS, FIG4_IN_BAR_LABELS, FIG4_XVALS, MATLAB_BAR_COLORS_FIG4, [0.5] * 12, "fig4a_mb_task_stream_assignment_source_style_demo", "Fig 4A MB task-stream assignment", "Task-stream assignment [%]")
- plot_matlab_like_three_roi(csvs["fig4b_mb_task_functional_correspondence_points"], csvs["fig4b_mb_task_functional_correspondence_noise_ceiling"], FIG4_MODEL_ORDER, FIG4_MODEL_LABELS, FIG4_IN_BAR_LABELS, FIG4_XVALS, MATLAB_BAR_COLORS_FIG4, [0.5] * 12, "fig4b_mb_task_functional_correspondence_source_style_demo", "Fig 4B MB task functional correspondence", "Functional similarity [r]")
- # %% [markdown]
- # ## Fig 5B: Spatial Weight Curves
- # %%
- def plot_fig5b_source_style(points, bands, stem_prefix, title_prefix, ylim_top):
- sns.set_theme(style="ticks")
- line_kwargs = {"marker": ".", "markersize": 18, "lw": 1, "markeredgecolor": "w", "markeredgewidth": 1.5}
- version_order = [v for v in ["self_supervised", "supervised"] if v in set(points["version"])]
- label_map = {"self_supervised": "Self-supervised", "supervised": "Supervised"}
- for streamx, stream in enumerate(CORE_ROI_NAMES):
- fig, ax = plt.subplots(figsize=(3, 5))
- df_stream = points[points["stream"] == stream].rename(columns={"spatial_weight": "Spatial_Weight", "corr": "Corr", "version": "Version"})
- sns.lineplot(
- data=df_stream,
- x="Spatial_Weight",
- y="Corr",
- hue="Version",
- hue_order=version_order,
- palette=[TDANN_COLORS[v] for v in version_order],
- errorbar="se",
- ax=ax,
- **line_kwargs,
- )
- ax.set_xscale("symlog", linthresh=0.09)
- ax.set_xlim([-0.01, 50])
- ax.set_xticks([], minor=True)
- ax.set_xticks([0, 0.1, 0.25, 0.5, 1.25, 2.5, 25])
- ax.set_xticklabels([0, 0.1, "", "", 1.25, "", 25])
- band = bands[bands["stream"] == stream].iloc[0]
- ax.axhspan(band["mean"] - band["std"], band["mean"] + band["std"], xmin=0, xmax=1, color="lightgray", alpha=0.75)
- if streamx == 2:
- h, _ = ax.get_legend_handles_labels()
- ax.legend(h, [label_map[v] for v in version_order], frameon=False)
- else:
- ax.legend([], [], frameon=False)
- ax.set_xlabel("Spatial Weight", fontsize=14)
- ax.set_ylabel("Correlation", fontsize=14)
- ax.set_yticks([])
- ax.spines["left"].set_visible(False)
- ax.spines["right"].set_visible(False)
- ax.spines["top"].set_visible(False)
- ax.set_ylim(bottom=0.0, top=ylim_top)
- ax.set_title(f"{title_prefix}: {stream}")
- savefig(fig, f"{stem_prefix}_{stream.lower()}_source_style_demo")
- plot_fig5b_source_style(csvs["fig5b_sc_spatial_correspondence_by_weight_points"], csvs["fig5b_sc_spatial_correspondence_brain_band_mean_sd"], "fig5b_top_spatial_correspondence", "Fig 5B top", 0.3)
- plot_fig5b_source_style(csvs["fig5b_sc_functional_correspondence_by_weight_points"], csvs["fig5b_sc_functional_correspondence_brain_band_mean_sd"], "fig5b_bottom_functional_correspondence", "Fig 5B bottom", 0.5)
- # %% [markdown]
- # ## Fig 5C: Effective Dimensionality vs Functional Similarity
- # %%
- points = csvs["fig5c_effective_dimensionality_vs_functional_similarity_points"].copy()
- points = points.rename(columns={"spatial_weight": "Spatial Weight", "type": "Type", "combined": "Combined", "stream": "Stream", "ed": "ED", "corr": "Corr"})
- bands = csvs["fig5c_brain_reference_bands_mean_sd"]
- for ridx, roi in enumerate(CORE_ROI_NAMES):
- fig, ax = plt.subplots(figsize=(7, 9))
- sns.set_theme(style="ticks")
- df_roi = points[points["Stream"] == roi]
- band = bands[bands["stream"] == roi].iloc[0]
- ax.axhspan(band["mean_corr"] - band["std_corr"], band["mean_corr"] + band["std_corr"], xmin=0, xmax=1, color="lightgray", alpha=0.75)
- ax.axvspan(band["mean_ed"] - band["std_ed"], band["mean_ed"] + band["std_ed"], color="lightgray", alpha=0.75)
- sup_open = df_roi[(df_roi["Type"] == "supervised") & ~(df_roi["Spatial Weight"].astype(str).isin(["2.5"]))]
- sim_open = df_roi[(df_roi["Type"] == "simCLR") & ~(df_roi["Spatial Weight"].astype(str).isin(["0.25", "0.5"]))]
- sup_filled = df_roi[(df_roi["Type"] == "supervised") & (df_roi["Spatial Weight"].astype(str).isin(["2.5"]))]
- sim_filled = df_roi[(df_roi["Type"] == "simCLR") & (df_roi["Spatial Weight"].astype(str).isin(["0.25", "0.5"]))]
- ax.scatter(sup_open["ED"], sup_open["Corr"], s=70, facecolors="white", edgecolors="#CB6D4A", linewidths=2, alpha=0.3, zorder=10, label="supervised")
- ax.scatter(sim_open["ED"], sim_open["Corr"], s=70, facecolors="white", edgecolors="#720298", linewidths=2, alpha=0.3, zorder=10, label="simCLR")
- ax.scatter(sup_filled["ED"], sup_filled["Corr"], s=130, color="#CB6D4A", alpha=0.65, zorder=11)
- ax.scatter(sim_filled["ED"], sim_filled["Corr"], s=90, color="#720298", alpha=0.65, zorder=11)
- ax.set_xscale("log")
- ax.set_ylim(0, 0.43)
- ax.set_xlabel("Effective Dimensionality")
- ax.set_ylabel("Correlation")
- ax.set_title(f"Fig 5C {roi}")
- ax.legend(frameon=False)
- ax.spines["right"].set_visible(False)
- ax.spines["top"].set_visible(False)
- savefig(fig, f"fig5c_ed_vs_functional_similarity_{roi.lower()}_source_style_demo")
- # %% [markdown]
- # ## Figs 6A/6B: Transfer Panels
- # %%
- def plot_transfer_source_style(points, value_col, stem, title, ylim=None):
- sns.set_theme(style="ticks")
- fig, ax = plt.subplots(figsize=(3, 6))
- raw_streams = {"Ventral": "Ventral", "Dorsal": "Parietal"}
- wide = points.pivot_table(index="line_index", columns="stream_display", values=value_col, aggfunc="first")
- for _, row in wide.dropna(subset=["Dorsal", "Ventral"]).iterrows():
- ax.plot([0.025, 0.975], [row["Dorsal"], row["Ventral"]], c="k", alpha=0.15)
- for stream_display, xpos, color in [("Ventral", 1, "#8C1A4C"), ("Dorsal", 0, "#377E2C")]:
- for hemi, marker in [("lh", "^"), ("rh", "o")]:
- vals = points[(points["stream_display"] == stream_display) & (points["hemi"] == hemi)][value_col]
- ax.scatter(
- np.full(len(vals), xpos), vals, color=color, s=70, alpha=0.95,
- edgecolors="w", marker=marker, linewidths=0.5,
- )
- ax.set_xticks([0, 1])
- ax.set_xticklabels(["Dorsal", "Ventral"], fontsize=16)
- ax.set_xlim([-0.2, 1.2])
- if ylim:
- ax.set_ylim(ylim)
- ax.spines["right"].set_visible(False)
- ax.spines["top"].set_visible(False)
- ax.set_ylabel(value_col.replace("_", " "))
- ax.set_title(title)
- savefig(fig, stem)
- plot_transfer_source_style(csvs["fig6a_object_position_transfer_points"], "accuracy_proportion", "fig6a_object_position_transfer_source_style_demo", "Fig 6A object position", ylim=(0.425, 0.535))
- plot_transfer_source_style(csvs["fig6b_imagenet_category_transfer_points"], "max_accuracy_percent", "fig6b_imagenet_category_transfer_source_style_demo", "Fig 6B ImageNet category")
- # %% [markdown]
- # ## Fig 6C: Model fLoc Selectivity
- # %%
- plot_floc_bar_strip(
- csvs["fig6c_model_floc_selectivity_subject_bars"],
- csvs["fig6c_model_floc_selectivity_hemi_dots"],
- "fig6c_model_floc_selectivity_source_style_demo",
- "Fig 6C model fLoc selectivity",
- hue_order=["Dorsal", "Lateral", "Ventral"],
- )
- # %% [markdown]
- # ## Fig 6D: Face-Selective RF Eccentricity
- # %%
- df = csvs["fig6d_face_selective_rf_eccentricity_plot_points"].copy()
- sns.set_theme(style="ticks")
- fig, ax = plt.subplots(figsize=(4, 12))
- order = ["Lateral", "Ventral"]
- palette = [ROI_COLORS[1], ROI_COLORS[2]]
- sns.stripplot(
- x="stream", y="eccen", hue="stream", jitter=0.1, linewidth=0.75, edgecolor="w",
- palette=palette, data=df[(df["hemi"] == "rh") & (df["stream"].isin(order))],
- order=order, dodge=True, size=9, ax=ax,
- )
- sns.stripplot(
- x="stream", y="eccen", hue="stream", jitter=0.2, linewidth=1, edgecolor="w", marker="^",
- palette=palette, data=df[(df["hemi"] == "lh") & (df["stream"].isin(order))],
- order=order, dodge=True, size=9, ax=ax,
- )
- sns.violinplot(
- x="stream", y="eccen", hue="stream", fill=True, linewidth=3, inner="box",
- saturation=0.9, palette=palette, data=df[df["stream"].isin(order)], order=order,
- dodge=False, ax=ax,
- )
- for collection in ax.collections:
- collection.set_alpha(0.85)
- ax.spines["right"].set_visible(False)
- ax.spines["top"].set_visible(False)
- ax.legend([], [], frameon=False)
- ax.set_xlabel("")
- ax.set_ylabel("RF eccentricity")
- ax.set_title("Fig 6D face-selective RF eccentricity")
- savefig(fig, "fig6d_face_selective_rf_eccentricity_source_style_demo")
- # %% [markdown]
- # ## S2A/S2B: Stream Correlations And Cross-ROI Fits
- # %%
- def plot_s2a(points, summary):
- data = points.copy()
- summary = summary.copy()
- pair_order = ["Ventral/Lateral", "Ventral/Parietal", "Lateral/Parietal"]
- comparison_order = ["Within", "Between"]
- colors = {"Within": "#bdbdbd", "Between": "#525252"}
- x = np.arange(len(pair_order))
- width = 0.34
- offsets = {"Within": -width / 2, "Between": width / 2}
- fig, ax = plt.subplots(figsize=(4.6, 3.4), facecolor="white")
- for comparison in comparison_order:
- vals = []
- errs = []
- for pair in pair_order:
- row = summary[(summary["stream_pair"] == pair) & (summary["comparison"] == comparison)].iloc[0]
- vals.append(float(row["mean_fisher_z"]))
- errs.append(float(row["sem_fisher_z"]))
- xpos = x + offsets[comparison]
- ax.bar(xpos, vals, yerr=errs, width=width, color=colors[comparison], edgecolor="black", linewidth=0.8, capsize=2, label=comparison)
- for idx, pair in enumerate(pair_order):
- subj_vals = data[(data["stream_pair"] == pair) & (data["comparison"] == comparison)]["fisher_z"].to_numpy(float)
- jitter = np.linspace(-0.06, 0.06, len(subj_vals)) if len(subj_vals) else []
- ax.scatter(np.full(len(subj_vals), xpos[idx]) + jitter, subj_vals, s=18, color="lightgray", edgecolor="black", linewidth=0.35, zorder=3)
- ax.set_xticks(x)
- ax.set_xticklabels(pair_order, rotation=25, ha="right")
- ax.set_ylabel("RSM correlation (Fisher z)")
- ax.legend(frameon=False)
- despine(ax)
- fig.tight_layout()
- savefig(fig, "s2a_within_between_stream_correlations_demo")
- def plot_s2b(points, summary):
- data = points.copy()
- summary = summary.copy()
- roi_order = ["Ventral", "Lateral", "Parietal"]
- model_order = ["same_roi", "other_roi1", "other_roi2"]
- model_labels = {"same_roi": "Same ROI", "other_roi1": "Other ROI 1", "other_roi2": "Other ROI 2"}
- colors = {"same_roi": "#f0f0f0", "other_roi1": "#969696", "other_roi2": "#252525"}
- x = np.arange(len(roi_order))
- width = 0.22
- offsets = {"same_roi": -width, "other_roi1": 0, "other_roi2": width}
- fig, ax = plt.subplots(figsize=(4.5, 4.4), facecolor="white")
- for model in model_order:
- vals = []
- errs = []
- for roi in roi_order:
- row = summary[(summary["roi"] == roi) & (summary["Model"] == model)].iloc[0]
- vals.append(float(row["mean_corrected"]))
- errs.append(float(row["sem_corrected"]))
- xpos = x + offsets[model]
- ax.bar(xpos, vals, yerr=errs, width=width, color=colors[model], edgecolor="black", linewidth=0.8, capsize=2, label=model_labels[model])
- for idx, roi in enumerate(roi_order):
- subj_vals = data[(data["roi"] == roi) & (data["Model"] == model)]["corrected"].to_numpy(float)
- jitter = np.linspace(-0.035, 0.035, len(subj_vals)) if len(subj_vals) else []
- ax.scatter(np.full(len(subj_vals), xpos[idx]) + jitter, subj_vals, s=18, color="lightgray", edgecolor="blue", linewidth=0.35, zorder=3)
- ax.axhline(1, color="black", linestyle="--", linewidth=1.0)
- ax.set_ylim(0, 1.05)
- ax.set_xticks(x)
- ax.set_xticklabels(roi_order, rotation=25, ha="right")
- ax.set_ylabel("Corrected R-squared")
- ax.legend(frameon=False, loc="upper right")
- despine(ax)
- fig.tight_layout()
- savefig(fig, "s2b_cross_roi_subject_to_subject_demo")
- plot_s2a(sheets["s2a_points"], sheets["s2a_summary"])
- plot_s2b(sheets["s2b_points"], sheets["s2b_summary"])
- # %% [markdown]
- # ## S3C: Voxel-To-Voxel Accuracy, Checkpoint 0
- # %%
- df = sheets["s3c_points"].copy()
- df["display_roi"] = df["roi"].replace({"Parietal": "Dorsal"})
- order = ["Dorsal", "Lateral", "Ventral"]
- fig, ax = plt.subplots(figsize=(3.2, 4.2), facecolor="white")
- x = np.arange(len(order))
- for idx, roi in enumerate(order):
- vals = df.loc[df["display_roi"] == roi, "accuracy_percent"].to_numpy(float)
- ax.bar(idx, np.nanmean(vals), yerr=np.nanstd(vals, ddof=1), width=0.62,
- color=STREAM_COLORS[roi], alpha=0.75, edgecolor="white", linewidth=0.8)
- jitter = np.linspace(-0.12, 0.12, len(vals)) if len(vals) else []
- ax.scatter(np.full(len(vals), idx) + jitter, vals, s=26, color="lightgray", edgecolor="black", linewidth=0.4, zorder=3)
- ax.axhline(33.3, color="black", linestyle="--", linewidth=1.0)
- ax.set_xticks(x)
- ax.set_xticklabels(order, rotation=30, ha="right")
- ax.set_ylabel("Voxel assignment correspondence (%)")
- ax.set_ylim(0, 100)
- despine(ax)
- fig.tight_layout()
- savefig(fig, "s3c_voxel2voxel_accuracy_checkpoint0_demo")
- # %% [markdown]
- # ## S5A/S5B: TDANN sw0.25 vs V1-Control
- # %%
- def plot_s5_paired(points, stem, ylabel):
- data = points.copy()
- data["ROI"] = data["ROI"].replace({"Parietal": "Dorsal"})
- subj = data.groupby(["condition", "subject", "ROI"], as_index=False)["result"].mean()
- cond_order = ["TDANN_sw0.25", "TDANN_sw0.25_v1_control"]
- cond_labels = ["TDANN\nsw0.25", "V1-control\nsw0.25"]
- fig, axes = plt.subplots(1, 3, figsize=(7.2, 3.3), sharey=True, facecolor="white")
- for ax, roi in zip(axes, STREAM_ORDER):
- wide = subj[subj["ROI"] == roi].pivot(index="subject", columns="condition", values="result").dropna()
- vals = [wide[c].to_numpy(float) for c in cond_order]
- means = [np.nanmean(v) for v in vals]
- sems = [np.nanstd(v, ddof=1) / np.sqrt(np.sum(np.isfinite(v))) for v in vals]
- ax.bar([0, 1], means, yerr=sems, width=0.62, color=["#d7d7d7", "#9e9e9e"],
- edgecolor="#595959", linewidth=0.9, capsize=2, zorder=1)
- for _, row in wide.iterrows():
- ax.plot([0, 1], [row[cond_order[0]], row[cond_order[1]]], color="black", alpha=0.22, linewidth=0.7, zorder=2)
- ax.scatter([0, 1], [row[cond_order[0]], row[cond_order[1]]], color=STREAM_COLORS[roi],
- edgecolor="white", linewidth=0.4, s=24, zorder=3)
- ax.set_title(roi)
- ax.set_xticks([0, 1])
- ax.set_xticklabels(cond_labels, rotation=25, ha="right")
- despine(ax)
- axes[0].set_ylabel(ylabel)
- fig.tight_layout()
- savefig(fig, stem)
- plot_s5_paired(sheets["s5_spat_points"], "s5a_tdann_v1_control_spatial_demo", "Spatial similarity [r]")
- plot_s5_paired(sheets["s5_func_points"], "s5b_tdann_v1_control_functional_demo", "Noise-corrected correlation")
- # %% [markdown]
- # ## S6A/S6B: ViT-Control Correspondence Panels
- # %%
- S6A_ORDER = [
- "MB_RN50_v2_detection",
- "MB_RN50_v2_vit_control_detection",
- "MB_RN50_v2_clip",
- "MB_RN50_v2_vit_control_clip",
- "MB_RN50_v2_categorization",
- "MB_RN50_v2_vit_control_categorization",
- ]
- S6B_ORDER = [
- "MB_RN50_v2_detection",
- "MB_RN50_v2_detection_vit_control",
- "MB_RN50_v2_clip",
- "MB_RN50_v2_clip_vit_control",
- "MB_RN50_v2_categorization",
- "MB_RN50_v2_categorization_vit_control",
- ]
- MODEL_LABELS = {
- "MB_RN50_v2_detection": "Det.",
- "MB_RN50_v2_vit_control_detection": "Det.\nV1",
- "MB_RN50_v2_detection_vit_control": "Det.\nV1",
- "MB_RN50_v2_clip": "CLIP",
- "MB_RN50_v2_vit_control_clip": "CLIP\nV1",
- "MB_RN50_v2_clip_vit_control": "CLIP\nV1",
- "MB_RN50_v2_categorization": "Cat.",
- "MB_RN50_v2_vit_control_categorization": "Cat.\nV1",
- "MB_RN50_v2_categorization_vit_control": "Cat.\nV1",
- }
- MODEL_COLORS = ["#80b380", "#80b380", "#6666cc", "#6666cc", "#cc8066", "#cc8066"]
- MODEL_ALPHA = [1.0, 0.48, 1.0, 0.48, 1.0, 0.48]
- def plot_s6(points, noise, model_order, stem, ylabel, chance=None, ylim=None):
- data = points.copy()
- data["ROIS"] = data["ROIS"].replace({"Parietal": "Dorsal"})
- noise = noise.copy()
- noise["ROI"] = noise["ROI"].replace({"Parietal": "Dorsal"})
- fig, axes = plt.subplots(1, 3, figsize=(9.0, 3.2), sharey=True, facecolor="white")
- x = np.array([0, 1, 2.15, 3.15, 4.3, 5.3])
- for ax, roi in zip(axes, STREAM_ORDER):
- sub = data[data["ROIS"] == roi]
- for idx, model in enumerate(model_order):
- vals = sub.loc[sub["model_type"] == model, "result"].to_numpy(float)
- ax.bar(x[idx], np.nanmean(vals), yerr=np.nanstd(vals, ddof=1), width=0.82,
- color=MODEL_COLORS[idx], alpha=MODEL_ALPHA[idx], edgecolor="none", capsize=1.8, zorder=1)
- lh = sub[(sub["model_type"] == model) & (sub["hemi"] == "lh")]["result"].to_numpy(float)
- rh = sub[(sub["model_type"] == model) & (sub["hemi"] == "rh")]["result"].to_numpy(float)
- ax.scatter(np.full(len(lh), x[idx]) - 0.08, lh, marker="^", s=11, color="none", edgecolor="black", linewidth=0.35, zorder=3)
- ax.scatter(np.full(len(rh), x[idx]) + 0.08, rh, marker="o", s=11, color="none", edgecolor="black", linewidth=0.35, zorder=3)
- nvals = noise.loc[noise["ROI"] == roi, "result"].to_numpy(float)
- if len(nvals):
- ax.axhspan(np.nanmean(nvals) - np.nanstd(nvals, ddof=1), np.nanmean(nvals) + np.nanstd(nvals, ddof=1),
- color="lightgray", alpha=0.35, zorder=0)
- if chance is not None:
- ax.axhline(chance, color="black", linestyle=":", linewidth=1.0)
- ax.set_title(roi)
- ax.set_xticks(x)
- ax.set_xticklabels([MODEL_LABELS[m] for m in model_order], rotation=0)
- if ylim:
- ax.set_ylim(*ylim)
- despine(ax)
- axes[0].set_ylabel(ylabel)
- fig.tight_layout()
- savefig(fig, stem)
- plot_s6(sheets["s6a_points"], sheets["s6a_noise"], S6A_ORDER, "s6a_vit_control_stream_assignment_demo", "Stream assignment (%)", chance=33, ylim=(0, 100))
- plot_s6(sheets["s6b_points"], sheets["s6b_noise"], S6B_ORDER, "s6b_vit_control_functional_correspondence_demo", "Functional correspondence [r]", ylim=(0, 0.36))
- # %% [markdown]
- # ## S10: TDANN CKA Across Spatial Weights
- # %%
- summary = sheets["s10_summary"].copy()
- layer_order = [
- "base_model.maxpool", "base_model.layer1.0", "base_model.layer1.1",
- "base_model.layer2.0", "base_model.layer2.1", "base_model.layer3.0", "base_model.layer3.1",
- "base_model.layer4.0", "base_model.layer4.1",
- ]
- pretty = [l.replace("base_model.", "").replace("layer", "layer ") for l in layer_order]
- ceiling = summary[summary["comparison"] == "ceiling pooled (sw0.0 + sw0.25)"].set_index("layer").reindex(layer_order)
- cross = summary[summary["comparison"] == "cross-weight (sw0.0 vs sw0.25)"].set_index("layer").reindex(layer_order)
- x = np.arange(len(layer_order))
- fig, ax = plt.subplots(figsize=(5.7, 2.8), facecolor="white")
- c_mean = ceiling["mean_cka"].to_numpy(float)
- c_sem = ceiling["sem_cka"].to_numpy(float)
- x_mean = cross["mean_cka"].to_numpy(float)
- ax.plot(x, c_mean, color="#2F6DAE", lw=1.4, label="Noise ceiling")
- ax.fill_between(x, c_mean - c_sem, c_mean + c_sem, color="#2F6DAE", alpha=0.18, linewidth=0)
- ax.plot(x, x_mean, color="#C55A11", marker="o", ms=4, lw=2.0, label="Cross-weight")
- ax.set_xticks(x)
- ax.set_xticklabels(pretty, rotation=35, ha="right")
- ax.set_ylabel("Linear CKA")
- ax.set_xlabel("TDANN layer")
- ax.set_ylim(0, 1)
- ax.grid(axis="y", color="#d9d9d9", linewidth=0.7, alpha=0.7)
- ax.legend(frameon=False, loc="lower left")
- despine(ax)
- fig.tight_layout()
- savefig(fig, "s10_cka_tdann_spatial_weight_demo")
- # %% [markdown]
- # ## S12A: Functional Correspondence Across Spatial Weights
- # %%
- def plot_s12a(points, bands):
- data = points.copy()
- data["stream"] = data["stream"].replace({"Parietal": "Dorsal"})
- bands = bands.copy()
- bands["stream"] = bands["stream"].replace({"Parietal": "Dorsal"})
- data["spatial_weight"] = pd.to_numeric(data["spatial_weight"])
- versions = [v for v in ["self_supervised", "supervised"] if v in set(data["version"])]
- palette = {"self_supervised": "#720298", "supervised": "#B59410"}
- labels = {"self_supervised": "SimCLR", "supervised": "Supervised"}
- fig, axes = plt.subplots(1, 3, figsize=(8.4, 3.0), sharey=True, facecolor="white")
- for ax, stream in zip(axes, STREAM_ORDER):
- sub = data[data["stream"] == stream]
- for version in versions:
- vsub = sub[sub["version"] == version]
- grouped = vsub.groupby("spatial_weight")["corr"].agg(["mean", "std", "count"]).reset_index().sort_values("spatial_weight")
- grouped["sem"] = grouped["std"] / np.sqrt(grouped["count"])
- ax.plot(grouped["spatial_weight"], grouped["mean"], marker=".", markersize=9, lw=1.5,
- color=palette.get(version, "black"), label=labels.get(version, version))
- ax.fill_between(grouped["spatial_weight"].to_numpy(float),
- (grouped["mean"] - grouped["sem"]).to_numpy(float),
- (grouped["mean"] + grouped["sem"]).to_numpy(float),
- color=palette.get(version, "black"), alpha=0.12, linewidth=0)
- band = bands[bands["stream"] == stream]
- if len(band):
- mean = float(band["mean"].iloc[0])
- sd = float(band["std"].iloc[0])
- ax.axhspan(mean - sd, mean + sd, color="lightgray", alpha=0.45, zorder=0)
- ax.set_xscale("symlog", linthresh=0.09)
- ax.set_xlim(-0.01, 50)
- ax.set_xticks([0, 0.1, 0.25, 0.5, 1.25, 2.5, 25])
- ax.set_xticklabels(["0", "0.1", "", "", "1.25", "", "25"])
- ax.set_title(stream)
- ax.set_xlabel("Spatial weight")
- despine(ax)
- axes[0].set_ylabel("Functional correspondence [r]")
- axes[-1].legend(frameon=False, loc="upper right")
- fig.tight_layout()
- savefig(fig, "s12a_functional_correspondence_by_weight_demo")
- plot_s12a(sheets["s12a_points"], sheets["s12a_brain_band"])
- # %% [markdown]
- # ## S11: HVM Transfer, Selected Tasks
- # %%
- def plot_s11(seedavg):
- data = seedavg.copy()
- data["spatial_weight"] = pd.to_numeric(data["spatial_weight"])
- task_order = ["categorization", "combined_position", "rotation_xz", "size"]
- task_titles = {
- "categorization": "Categorization",
- "combined_position": "Position",
- "rotation_xz": "Rotation XZ",
- "size": "Size",
- }
- ylabels = {
- "categorization": "ImageNet accuracy (%)",
- "combined_position": "Position accuracy",
- "rotation_xz": "Rotation correlation [r]",
- "size": "Size correlation [r]",
- }
- fig, axes = plt.subplots(1, 4, figsize=(11.5, 3.0), facecolor="white")
- for ax, task in zip(axes, task_order):
- sub = data[data["task"] == task]
- for stream in STREAM_ORDER:
- grouped = (
- sub[sub["stream"] == stream]
- .groupby("spatial_weight")["result"]
- .agg(["mean", "std", "count"])
- .reset_index()
- .sort_values("spatial_weight")
- )
- grouped["sem"] = grouped["std"] / np.sqrt(grouped["count"])
- ax.plot(
- grouped["spatial_weight"],
- grouped["mean"],
- marker=".",
- markersize=9,
- lw=1.6,
- color=STREAM_COLORS[stream],
- label=stream,
- )
- ax.fill_between(
- grouped["spatial_weight"].to_numpy(float),
- (grouped["mean"] - grouped["sem"]).to_numpy(float),
- (grouped["mean"] + grouped["sem"]).to_numpy(float),
- color=STREAM_COLORS[stream],
- alpha=0.12,
- linewidth=0,
- )
- ax.set_xscale("symlog", linthresh=0.09)
- ax.set_xlim(-0.01, 50)
- ax.set_xticks([0, 0.1, 0.25, 0.5, 1.25, 2.5, 25])
- ax.set_xticklabels(["0", "0.1", "", "", "1.25", "", "25"])
- ax.set_title(task_titles[task])
- ax.set_xlabel("Spatial weight")
- ax.set_ylabel(ylabels[task])
- despine(ax)
- axes[-1].legend(frameon=False, loc="best")
- fig.tight_layout()
- savefig(fig, "s11_hvm_transfer_selected_tasks_demo")
- plot_s11(sheets["s11_seedavg"])
- # %% [markdown]
- # ## S12B: Supervised vs SimCLR Across Spatial Weights
- # %%
- def plot_s12b(points, bands):
- data = points.copy()
- data["spatial_weight"] = pd.to_numeric(data["spatial_weight"])
- bands = bands.copy()
- palette = {"simCLR": "#720298", "supervised": "#B59410"}
- labels = {"simCLR": "SimCLR", "supervised": "Supervised"}
- versions = [v for v in ["simCLR", "supervised"] if v in set(data["version"])]
- fig, axes = plt.subplots(1, 3, figsize=(8.4, 3.0), sharey=True, facecolor="white")
- for ax, stream in zip(axes, STREAM_ORDER):
- sub = data[data["stream"] == stream]
- for version in versions:
- grouped = (
- sub[sub["version"] == version]
- .groupby("spatial_weight")["means"]
- .agg(["mean", "std", "count"])
- .reset_index()
- .sort_values("spatial_weight")
- )
- grouped["sem"] = grouped["std"] / np.sqrt(grouped["count"])
- ax.plot(
- grouped["spatial_weight"],
- grouped["mean"],
- marker=".",
- markersize=9,
- lw=1.5,
- color=palette[version],
- label=labels[version],
- )
- ax.fill_between(
- grouped["spatial_weight"].to_numpy(float),
- (grouped["mean"] - grouped["sem"]).to_numpy(float),
- (grouped["mean"] + grouped["sem"]).to_numpy(float),
- color=palette[version],
- alpha=0.12,
- linewidth=0,
- )
- band = bands[bands["stream"] == stream]
- if len(band):
- mean = float(band["mean"].iloc[0])
- sd = float(band["std"].iloc[0])
- ax.axhspan(mean - sd, mean + sd, color="lightgray", alpha=0.45, zorder=0)
- ax.set_xscale("symlog", linthresh=0.09)
- ax.set_xlim(-0.01, 50)
- ax.set_xticks([0, 0.1, 0.25, 0.5, 1.25, 2.5, 25])
- ax.set_xticklabels(["0", "0.1", "", "", "1.25", "", "25"])
- ax.set_title(stream)
- ax.set_xlabel("Spatial weight")
- despine(ax)
- axes[0].set_ylabel("Noise-corrected correlation")
- axes[-1].legend(frameon=False, loc="upper right")
- fig.tight_layout()
- savefig(fig, "s12b_supervised_vs_simclr_by_weight_demo")
- plot_s12b(sheets["s12b_points"], sheets["s12b_brain_band"])
- # %%
- print(f"Wrote combined demo plots to {OUT}")
- print("PNG files:", len(list(OUT.glob("*.png"))))
- print("PDF files:", len(list(OUT.glob("*.pdf"))))
demo_plot_combined_source_data.ipynb at commit d85997e, under Apache-2.0 · at the source
Overview
- Department of Psychology, Stanford University, Stanford, CA USA
- Department of Computer Science, Stanford University, Stanford, CA USA
- Neurosciences Graduate Program, Stanford University, Stanford, CA USA
- Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, Minneapolis, MN USA
- Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA USA
Abstract
The abstract is not reproduced here: the paper's license (CC BY-NC-ND) does not allow it. Read it in the paper, at the publisher or on Europe PMC.
Repositories
Its files are read in the Code ↔ Paper reader above, with 16 matches between paragraphs and lines of code.
dawnfinzi/spacestream
d85997e327886f119441cc6a8355584373878d29, 26 September 2026Availability: 1 check, the latest on 27 September 2026: the link answers
- 27 September 2026: the link answers
72 files
- matlab/
F03_B.m , MATLAB, 231 lines - matlab/
F03_C.m , MATLAB, 223 lines - matlab/
F04_A.m , MATLAB, 278 lines, 1 match - matlab/
F04_B.m , MATLAB, 275 lines - matlab/
create_distance_matrix.m , MATLAB, 78 lines - matlab/
prepare_betas.m , MATLAB, 131 lines - matlab/
read_main_text_source_ta , MATLAB, 12 linesble.m - matlab/
transform_selectivity_to , MATLAB, 25 lines, 1 match_fsaverage.m - matlab/
verify_distance_matrix.m , MATLAB, 44 lines - notebooks/
F01_A.ipynb , Jupyter, 152 lines - notebooks/
F01_B.ipynb , Jupyter, 254 lines - notebooks/
F03_B_processing.ipynb , Jupyter, 219 lines, 1 match - notebooks/
F03_C_processing.ipynb , Jupyter, 320 lines - notebooks/
F04_A_processing.ipynb , Jupyter, 194 lines, 1 match - notebooks/
F04_B_processing.ipynb , Jupyter, 173 lines - notebooks/
F05_B.ipynb , Jupyter, 408 lines - notebooks/
F05_C.ipynb , Jupyter, 302 lines - notebooks/
F06_A_B.ipynb , Jupyter, 307 lines - notebooks/
F06_C.ipynb , Jupyter, 354 lines - notebooks/
F06_D.ipynb , Jupyter, 191 lines - notebooks/
demo_plot_combined_sourc , Jupyter, 991 lines, 2 matchese_data.ipynb - scripts/
TDANN_optimal_thirds.py , Python, 157 lines - scripts/
calc_effective_dimension , Python, 146 linesality.py - scripts/
calc_smoothness.py , Python, 304 lines - scripts/
cka_mb18_overlap.py , Python, 224 lines, 1 match - scripts/
cka_tdann_spatial_weight , Python, 309 lines.py - scripts/
eval_robustness.py , Python, 258 lines - scripts/
fitting_one_to_one_unit2 , Python, 705 lines, 1 matchvoxel.py - scripts/
fitting_one_to_one_voxel , Python, 347 lines2voxel.py - scripts/
fitting_regression_brain , Python, 246 lines2brain_by_area.py - scripts/
fitting_regression_model , Python, 221 lines2brain_by_layer.py - scripts/
functional_swap.py , Python, 368 lines - scripts/
linear_eval.py , Python, 381 lines - scripts/
test_overall_transfer.py , Python, 127 lines - scripts/
test_transfer_by_stream. , Python, 135 linespy - setup.py, Python, 36 lines
- spacestream/
analyses/ , Python, 249 lines, 1 matchbase_fitter.py - spacestream/
analyses/ , Python, 15 lineseffective_dimensionality .py - spacestream/
analyses/ , Python, 271 lineshvm_fitter.py - spacestream/
analyses/ , Python, 74 linessmoothness.py - spacestream/
core/ , Python, 167 linesconstants.py - spacestream/
core/ , Python, 232 linesfeature_extractor.py - spacestream/
core/ , Python, 76 linesfeature_saver.py - spacestream/
core/ , Python, 175 linesfit.py - spacestream/
core/ , Python, 122 lineslosses.py - spacestream/
core/ , Python, 21 linespaths.py - spacestream/
core/ , Python, 159 linespositions.py - spacestream/
core/ , Python, 320 linesswapopt.py - spacestream/
datasets/ , Python, 115 lineshvm.py - spacestream/
datasets/ , Python, 162 linesimagenet.py - spacestream/
datasets/ , Python, 210 linesnsd.py - spacestream/
datasets/ , Python, 365 linessine_gratings.py - spacestream/
models/ , Python, 4 lines__init__.py - spacestream/
models/ , Python, 107 linesfast_resnet3d.py - spacestream/
models/ , Python, 340 linesresnet.py - spacestream/
models/ , Python, 228 linesresnet3d.py - spacestream/
models/ , Python, 116 lines, 1 matchslow_fast.py - spacestream/
models/ , Python, 56 linesslow_resnet3d.py - spacestream/
models/ , Python, 64 linesspatial_resnet.py - spacestream/
utils/ , Python, 106 linesarray_utils.py - spacestream/
utils/ , Python, 20 linesdataloader_utils.py - spacestream/
utils/ , Python, 95 linesgeneral_utils.py - spacestream/
utils/ , Python, 436 lines, 2 matchesget_utils.py - spacestream/
utils/ , Python, 328 lines, 1 matchmapping_utils.py - spacestream/
utils/ , Python, 60 linesmemory_utils.py - spacestream/
utils/ , Python, 140 linesmetric_utils.py - spacestream/
utils/ , Python, 65 linesplot_utils.py - spacestream/
utils/ , Python, 203 linesregression_utils.py - spacestream/
utils/ , Python, 74 linesslowfast_utils.py - spacestream/
utils/ , Python, 142 linesswapopt_utils.py - LICENSE, License, 201 lines
- README.md, Text, 45 lines
neuroailab/TDANN
80c585df69dcb831d1d6802e0776a8cf25d8cfef, 18 October 2024Availability: 1 check, the latest on 27 September 2026: the link answers
- 27 September 2026: the link answers
95 files
- configs/
__init__.py , Python, 1 line - demo/
demo.ipynb , Jupyter, 77 lines - demo/
get_imagenet_logits.ipyn , Jupyter, 130 linesb - demo/
src/ , Python, 1 line__init__.py - demo/
src/ , Python, 34 linesdata.py - demo/
src/ , Python, 140 linesfeatures.py - demo/
src/ , Python, 40 linesindexing.py - demo/
src/ , Python, 39 linesmodel.py - demo/
src/ , Python, 107 linespositions.py - scripts/
precompute_eigvals.py , Python, 112 lines - scripts/
precompute_wiring.py , Python, 167 lines - scripts/
run_all_neural_fits.py , Python, 53 lines - scripts/
run_brainscore_benchmark , Python, 102 liness.py - scripts/
save_features_from_confi , Python, 54 linesg.py - scripts/
save_retinotopic_positio , Python, 151 linesns.py - scripts/
spatial_loss_curves/ , Python, 82 linesby_spatial_loss.py - scripts/
spatial_loss_curves/ , Python, 80 linesby_task_loss.py - scripts/
supplement/ , Python, 123 lineslayer_selection.py - scripts/
supplement/ , Python, 82 linesmake_orientation_biased_ sine_grating_dir.py - scripts/
supplement/ , Python, 97 linessave_retinal_wave_featur es_from_config.py - scripts/
supplement/ , Python, 59 linessave_vonenet_features.py - scripts/
supplement/ , Python, 72 linestraining_curves/ spatial_loss_curves.py - scripts/
supplement/ , Python, 61 linestraining_curves/ task_loss_curves.py - scripts/
swapopt_from_config.py , Python, 86 lines - setup.py, Python, 3 lines
- spacetorch/
__init__.py , Python, 1 line - spacetorch/
analyses/ , Python, 1 line__init__.py - spacetorch/
analyses/ , Python, 66 linesalpha.py - spacetorch/
analyses/ , Python, 120 linescore.py - spacetorch/
analyses/ , Python, 19 linesdimensionality.py - spacetorch/
analyses/ , Python, 91 linesfloc.py - spacetorch/
analyses/ , Python, 43 linesgfb.py - spacetorch/
analyses/ , Python, 17 linesimagenet.py - spacetorch/
analyses/ , Python, 64 linesrsa.py - spacetorch/
analyses/ , Python, 138 linessine_gratings.py - spacetorch/
analyses/ , Python, 51 lineswiring_length.py - spacetorch/
bmi/ , Python, 1 line__init__.py - spacetorch/
bmi/ , Python, 112 linesopt.py - spacetorch/
bmi/ , Python, 76 linespattern.py - spacetorch/
colormaps.py , Python, 42 lines - spacetorch/
constants.py , Python, 29 lines - spacetorch/
datasets/ , Python, 59 lines, 1 match__init__.py - spacetorch/
datasets/ , Python, 200 lines, 1 matchfloc.py - spacetorch/
datasets/ , Python, 86 linesimagenet.py - spacetorch/
datasets/ , Python, 34 linesnoise.py - spacetorch/
datasets/ , Python, 43 linesnsd.py - spacetorch/
datasets/ , Python, 238 linesretinal_waves.py - spacetorch/
datasets/ , Python, 55 linesringach_2002.py - spacetorch/
datasets/ , Python, 188 linessine_gratings.py - spacetorch/
datasets/ , Python, 23 linestransforms.py - spacetorch/
feature_extractor.py , Python, 179 lines - spacetorch/
feature_saver.py , Python, 93 lines - spacetorch/
losses/ , Python, 3 lines__init__.py - spacetorch/
losses/ , Python, 58 lines, 1 matchcross_entropy_spatial_co rrelation_loss.py - spacetorch/
losses/ , Python, 119 lineslosses_numpy.py - spacetorch/
losses/ , Python, 110 lineslosses_torch.py - spacetorch/
losses/ , Python, 58 linesspatial_correlation_simc lr_info_nce_loss.py - spacetorch/
maps/ , Python, 242 lines__init__.py - spacetorch/
maps/ , Python, 267 linesit_map.py - spacetorch/
maps/ , Python, 310 linesnsd_floc.py - spacetorch/
maps/ , Python, 121 linespatch.py - spacetorch/
maps/ , Python, 121 linespinwheel_detector.py - spacetorch/
maps/ , Python, 430 linesscreenshot_maps.py - spacetorch/
maps/ , Python, 114 linessmoother.py - spacetorch/
maps/ , Python, 288 linessom.py - spacetorch/
maps/ , Python, 240 linesv1_map.py - spacetorch/
models/ , Python, 236 lines__init__.py - spacetorch/
models/ , Python, 18 linesheads/ identity.py - spacetorch/
models/ , Python, 143 linespositions.py - spacetorch/
models/ , Python, 250 linestrunks/ resnet.py - spacetorch/
models/ , Python, 476 linestrunks/ resnet_chanx.py - spacetorch/
paths.py , Python, 48 lines - spacetorch/
stimulation.py , Python, 134 lines - spacetorch/
swapopt.py , Python, 317 lines - spacetorch/
train_steps/ , Python, 124 linescustom_train_step.py - spacetorch/
types.py , Python, 35 lines - spacetorch/
utils/ , Python, 2 lines__init__.py - spacetorch/
utils/ , Python, 145 linesarray_utils.py - spacetorch/
utils/ , Python, 159 linesfigure_utils.py - spacetorch/
utils/ , Python, 97 linesgeneric_utils.py - spacetorch/
utils/ , Python, 21 linesgpu_utils.py - spacetorch/
utils/ , Python, 31 linesoptimization_utils.py - spacetorch/
utils/ , Python, 171 linesplot_utils.py - spacetorch/
utils/ , Python, 545 linesspatial_utils.py - spacetorch/
utils/ , Python, 112 linestorch_utils.py - spacetorch/
utils/ , Python, 1 linevissl/ __init__.py - spacetorch/
utils/ , Python, 63 linesvissl/ extras.py - spacetorch/
utils/ , Python, 310 linesvissl/ hooks.py - spacetorch/
utils/ , Python, 45 linesvissl/ performance.py - spacetorch/
utils/ , Python, 6 linesvissl/ registration.py - spacetorch/
utils/ , Python, 116 linesvissl/ spatial_loss_eval.py - spacetorch/
utils/ , Python, 118 linesvissl/ task_loss_eval.py - spacetorch/
wiring_length.py , Python, 239 lines - train.py, Python, 50 lines
- README.md, Text, 60 lines
Zenodo 20753397
Availability: 1 check, the latest on 27 September 2026: the link answers (HTTP 200)
- 27 September 2026: the link answers (HTTP 200)
72 files
- matlab/
F03_B.m , MATLAB, 231 lines - matlab/
F03_C.m , MATLAB, 223 lines - matlab/
F04_A.m , MATLAB, 278 lines - matlab/
F04_B.m , MATLAB, 275 lines - matlab/
create_distance_matrix.m , MATLAB, 78 lines - matlab/
prepare_betas.m , MATLAB, 131 lines - matlab/
read_main_text_source_ta , MATLAB, 12 linesble.m - matlab/
transform_selectivity_to , MATLAB, 25 lines_fsaverage.m - matlab/
verify_distance_matrix.m , MATLAB, 44 lines - notebooks/
F01_A.ipynb , Jupyter, 152 lines - notebooks/
F01_B.ipynb , Jupyter, 254 lines - notebooks/
F03_B_processing.ipynb , Jupyter, 219 lines - notebooks/
F03_C_processing.ipynb , Jupyter, 320 lines - notebooks/
F04_A_processing.ipynb , Jupyter, 194 lines - notebooks/
F04_B_processing.ipynb , Jupyter, 173 lines - notebooks/
F05_B.ipynb , Jupyter, 408 lines - notebooks/
F05_C.ipynb , Jupyter, 302 lines - notebooks/
F06_A_B.ipynb , Jupyter, 307 lines - notebooks/
F06_C.ipynb , Jupyter, 354 lines - notebooks/
F06_D.ipynb , Jupyter, 191 lines - notebooks/
demo_plot_combined_sourc , Jupyter, 991 linese_data.ipynb - scripts/
TDANN_optimal_thirds.py , Python, 157 lines - scripts/
calc_effective_dimension , Python, 146 linesality.py - scripts/
calc_smoothness.py , Python, 304 lines - scripts/
cka_mb18_overlap.py , Python, 224 lines - scripts/
cka_tdann_spatial_weight , Python, 309 lines.py - scripts/
eval_robustness.py , Python, 258 lines - scripts/
fitting_one_to_one_unit2 , Python, 705 linesvoxel.py - scripts/
fitting_one_to_one_voxel , Python, 347 lines2voxel.py - scripts/
fitting_regression_brain , Python, 246 lines2brain_by_area.py - scripts/
fitting_regression_model , Python, 221 lines2brain_by_layer.py - scripts/
functional_swap.py , Python, 368 lines - scripts/
linear_eval.py , Python, 381 lines - scripts/
test_overall_transfer.py , Python, 127 lines - scripts/
test_transfer_by_stream. , Python, 135 linespy - setup.py, Python, 36 lines
- spacestream/
analyses/ , Python, 249 linesbase_fitter.py - spacestream/
analyses/ , Python, 15 lineseffective_dimensionality .py - spacestream/
analyses/ , Python, 271 lineshvm_fitter.py - spacestream/
analyses/ , Python, 74 linessmoothness.py - spacestream/
core/ , Python, 167 linesconstants.py - spacestream/
core/ , Python, 232 linesfeature_extractor.py - spacestream/
core/ , Python, 76 linesfeature_saver.py - spacestream/
core/ , Python, 175 linesfit.py - spacestream/
core/ , Python, 122 lineslosses.py - spacestream/
core/ , Python, 21 linespaths.py - spacestream/
core/ , Python, 159 linespositions.py - spacestream/
core/ , Python, 320 linesswapopt.py - spacestream/
datasets/ , Python, 115 lineshvm.py - spacestream/
datasets/ , Python, 162 linesimagenet.py - spacestream/
datasets/ , Python, 210 linesnsd.py - spacestream/
datasets/ , Python, 365 linessine_gratings.py - spacestream/
models/ , Python, 4 lines__init__.py - spacestream/
models/ , Python, 107 linesfast_resnet3d.py - spacestream/
models/ , Python, 340 linesresnet.py - spacestream/
models/ , Python, 228 linesresnet3d.py - spacestream/
models/ , Python, 116 linesslow_fast.py - spacestream/
models/ , Python, 56 linesslow_resnet3d.py - spacestream/
models/ , Python, 64 linesspatial_resnet.py - spacestream/
utils/ , Python, 106 linesarray_utils.py - spacestream/
utils/ , Python, 20 linesdataloader_utils.py - spacestream/
utils/ , Python, 95 linesgeneral_utils.py - spacestream/
utils/ , Python, 436 linesget_utils.py - spacestream/
utils/ , Python, 328 linesmapping_utils.py - spacestream/
utils/ , Python, 60 linesmemory_utils.py - spacestream/
utils/ , Python, 140 linesmetric_utils.py - spacestream/
utils/ , Python, 65 linesplot_utils.py - spacestream/
utils/ , Python, 203 linesregression_utils.py - spacestream/
utils/ , Python, 74 linesslowfast_utils.py - spacestream/
utils/ , Python, 142 linesswapopt_utils.py - LICENSE, License, 201 lines
- README.md, Text, 36 lines
Code availability statement
The paper has a code availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:
- it points to the authors' code: dawnfinzi/
spacestream , neuroailab/TDANN
Read it in the paper: doi.org/10.1038/s41467-026-76098-y.
Tracing map
Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.
What the map holds:
- 3 repositories of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
- 234 scripts, each with its path and the digest of its content;
- 16 matches between paragraphs of the paper and lines of the code (method lexical-v1);
- neither the text of the paper nor the code itself.
Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.
Data
Datasets cited
Data availability statement
The paper has a data availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:
- it points to a dataset: OSF qy32x
Read it in the paper: doi.org/10.1038/s41467-026-76098-y.
Versions
The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.
Version 1, 27 September 2026: the first record
Recorded: type, language, journal, volume, issue, pages, dates, 5 authors, 3 keywords, 9 MeSH terms, 3 funders, 88 references.
Cite
This paper
Finzi, D., Margalit, E., Kay, K., Yamins, D. L. K., & Grill-Spector, K. (2026). A single computational objective can produce specialization of streams in visual cortex. Nature communications, 17(1), 9762. https://
BibTeX
@article{finzi2026single
author = {Finzi, Dawn and Margalit, Eshed and Kay, Kendrick and Yamins, Daniel L K and Grill-Spector, Kalanit},
title = {{A single computational objective can produce specialization of streams in visual cortex}},
journal = {Nature communications},
year = {2026},
month = aug,
volume = {17},
number = {1},
pages = {9762},
publisher = {Nature Publishing Group},
issn = {2041-1723},
doi = {10.1038/
url = {https://
pmid = {42733086},
pmcid = {PMC13572372}
}
RIS
TY - JOUR
AU - Finzi, Dawn
AU - Margalit, Eshed
AU - Kay, Kendrick
AU - Yamins, Daniel L K
AU - Grill-Spector, Kalanit
TI - A single computational objective can produce specialization of streams in visual cortex
T2 - Nature communications
J2 - Nat Commun
PY - 2026
DA - 2026/
VL - 17
IS - 1
SP - 9762
SN - 2041-1723
PB - Nature Publishing Group
DO - 10.1038/
UR - https://
LA - en
ER -
CSL-JSON
{
"id": "10.1038/
"type": "article-journal",
"title": "A single computational objective can produce specialization of streams in visual cortex",
"container-title": "Nature communications",
"author": [
{
"family": "Finzi",
"given": "Dawn"
},
{
"family": "Margalit",
"given": "Eshed"
},
{
"family": "Kay",
"given": "Kendrick"
},
{
"family": "Yamins",
"given": "Daniel L K"
},
{
"family": "Grill-Spector",
"given": "Kalanit"
}
],
"container-title-short":
"volume": "17",
"issue": "1",
"page": "9762",
"DOI": "10.1038/
"PMID": "42733086",
"PMCID": "PMC13572372",
"ISSN": "2041-1723",
"publisher": "Nature Publishing Group",
"URL": "https://
"language": "en",
"issued": {
"date-parts": [
[
2026,
8,
13
]
]
}
}
The tracing map gets a citation of its own once an author has validated it and it has a DOI.
Similar papers
The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.
- [1] doi:10.1038/s41467-024-53147-y [code]
- A large-scale examination of inductive biases shaping high-level visual representation in brains and machines.Journal: Nature communicationsIn common: h5py, Pillow, NiBabel, 8 other tools, 14 references, author Kendrick Kay
- [2] doi:10.1038/s42256-025-01072-0 [code]
- High-level visual representations in the human brain are aligned with large language models.Journal: Nature machine intelligenceIn common: scikit-image, h5py, Pillow, 10 other tools, 12 references, author Kendrick Kay
- [3] doi:10.1038/s42003-026-10169-0 [code]
- Shared representations in brains and models reveal a two-route cortical organization during scene perception.Journal: Communications biologyIn common: Hugging Face Transformers, h5py, Pillow, 9 other tools, 14 references
- [4] doi:10.1371/journal.pbio.3002366 [code]
- Many but not all deep neural network audio models capture brain responses and exhibit correspondence between model stages and brain regions.Journal: PLoS biologyIn common: Hugging Face Transformers, xarray, FreeSurfer, 11 other tools, 10 references
- [5] doi:10.1162/imag_a_00488 [code]
- Individual variation in the functional lateralization of human ventral temporal cortex: Local competition and long-range coupling.Journal: Imaging neuroscience (Cambridge, Mass.)In common: FreeSurfer, statsmodels, NiBabel, 6 other tools, 16 references
- [6] doi:10.1038/s42003-026-09685-w [code]
- Achieving more human brain-like vision via human EEG representational alignment.Journal: Communications biologyIn common: Hugging Face Transformers, xarray, h5py, 9 other tools, 11 references
- [7] doi:10.1038/s41467-026-69345-9 [code]
- A 7T fMRI dataset of synthetic images for out-of-distribution modeling of vision.Journal: Nature communicationsIn common: FreeSurfer, h5py, Pillow, 8 other tools, 11 references, author Kendrick Kay
- [8] doi:10.1038/s41467-026-70065-3 [code]
- Local lateral connectivity is sufficient for replicating cortex-like topographical organization in deep neural networks.Journal: Nature communicationsIn common: xarray, scikit-image, h5py, 8 other tools, 12 references
- [9] doi:10.1038/s41467-024-50310-3 [code]
- Modeling short visual events through the BOLD moments video fMRI dataset and metadata.Journal: Nature communicationsIn common: FreeSurfer, Pillow, statsmodels, 7 other tools, 11 references, author Kendrick Kay
- [10] doi:10.1038/s41562-025-02220-7 [code]
- End-to-end topographic networks as models of cortical map formation and human visual behaviour.Journal: Nature human behaviourIn common: h5py, statsmodels, seaborn, 5 other tools, 15 references
Contribute
The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.
Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.
Claim this paper
Correct its record
Say what each link of this record is, remove the ones that are not the paper's, add the ones that are missing. The correction becomes a new version of the record, in its Versions section.
Validate its tracing map
You validate the map as this page shows it: 3 repositories of the authors' code, each at its verified commit and with its license, 234 scripts, and 16 matches between paragraphs and code (see the Code and Map sections). It then receives a DOI on Zenodo, with you (your ORCID iD) and OSCR as its creators; the code itself is not deposited.
The map's fingerprint: sha256:26d71a27603f6f35…
Add the badge to the code's README
If you are an author or a maintainer of this paper's code, you can add this badge. It says what OSCR knows of the paper (its code, its validated tracing map) and links the code to this page. Paste one of these snippets under the README's title. Keep the two marks around it, so that the badge can be replaced in place and is never added twice. OSCR changes nothing in your repository.
Markdown
<!-- oscr-badge:start -->
[](https://openscicode.org/badge/doi/10.1038/s41467-026-76098-y)
<!-- oscr-badge:end -->reStructuredText
.. oscr-badge:start
.. image:: https://openscicode.org/badge/doi/10.1038/s41467-026-76098-y.svg
:target: https://openscicode.org/badge/doi/10.1038/s41467-026-76098-y
:alt: OSCR
.. oscr-badge:endHTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.1038/s41467-026-76098-y"><img src="https://openscicode.org/badge/doi/10.1038/s41467-026-76098-y.svg" alt="OSCR"></a>
<!-- oscr-badge:end -->On GitHub, open the README, choose Edit (the pencil), paste the snippet under the title, then "Commit changes".
If you choose to, you can have the badge kept up to date for you, with oscr badge --write from the command line or with the GitHub Action in your repository's workflows. Every other form and option (a card, AsciiDoc, LaTeX, shields.io) is on the badge page.
Request its removal
To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).
Discussion, reproductions, activity
Discussion: code errors and mismatches between the code and the paper about this paper and its code, from readers and its authors, beside the paper and its tracing map.
Report a mismatch between the code and the paper · Report a code error (signed in)
Reproductions: reports from readers who ran the authors' code and did not get the paper's result, what they ran, with which environment, commit and data.
Report a reproduction failure (signed in)
Versions of its code: the releases of its code tied to a version of this paper (the preprint, the accepted manuscript, the version of record…), each with the tracing map versioned with it.
Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.
