How to train your sponge: Imprinting material memories in marine sponge tissue
Abstract
Sponges, though relatively simple in structure and function, are living animals and thus operate out of thermodynamic equilibrium. Often, they appear in aquatic environments with a moving current. It is natural to assume that they can adapt, actively or passively, to such an external drive. Here, we hypothesize that sponge tissue encodes material memories from cyclic driving that are similar to what they experience from moving current in vivo. We simulate externally driven current by way of oscillatory shear strain applied by a rheometer and show that sponge tissue remembers aspects of its previous loading history. In particular, we uncover that sponges, both living and decellularized dead skeletons, can form memories of two different kinds: memory of the largest strain, as well as memories of the training amplitude. We find that sponges with siliceous spicules embedded in their tissue have a higher capacity for both kinds of memory when compared to a species with no spicules, but surprisingly, a synthetic cleaning sponge also outperformed the species without spicules. We also shed light on so-called orthogonal memory effects. Our results broaden the class of disordered systems that form memories to include spongy materials.
This readme file was generated on 2025-12-30 and edited on 2026-02-15 by E.A.K.
GENERAL INFORMATION
Title of Dataset: Data from "How to train your sponge: imprinting material memories in marine sponge tissue"
Date of data collection: Jul--Oct 2025
Geographic location of data collection: Syracuse, NY
Information about funding sources that supported the collection of the data: None
DATA & FILE OVERVIEW
General Description:
This dataset has five main folders inside 'data.zip', one for each of the five materials studied: 'Aplysina fulva', 'Cliona celata', 'Halichondria bowerbanki', 'liquid PDMS', and 'synthetic sponge'. Within each of these folders are subfolders for each different rheological protocol carried out: 'memprotocol1' for the Mullins effect and 'memprotocol2' for constant amplitude annealing, as well as a folder labeled 'pics' or 'pix' with pictures of the materials. Each 'memprotocolx' folder also has a subfolder 'plots' with plots created by basic Python code. Each raw .csv file is labeled with the date acquired, the type of sample, which memory protocol it was, the frequency of oscillatory shear, and which number of sample it was. For 'memprotocol1' data, the files are also labeled with 'd' for whether it was a downwards amplitude sweep or 'u' for whether it was an upwards amplitude sweep. Each raw .csv file contains columns for the gap height (mm), action time (s), sequence time (s), angular displacement (rad), torque (N*m), normal force (N), angular frequency (rad/s), shear strain (%), shear stress (Pa), shear modulus (elastic component) (Pa), and shear modulus (viscous component) (Pa). There are also .csv files labeled by 'summary' at the end of the filename for summarized data for a given sample. Summarized data may also be found in the rows at the ends of the raw files. For Aplysina fulva and Halichondria bowerbanki, there are more samples listed than what was used in the paper. This was due to slippage in the rheometer plates or noise in the data.
METHODOLOGICAL INFORMATION
Methods for processing the data: numerical analysis in Python
DATA-SPECIFIC INFORMATION FOR: MEMPROTOCOL1 SUMMARIES
Number of variables: 10 (multiplied by the number of upwards and downwards strain sweeps completed, which was 4: up to 5%, up to 8%, up to 11%, and up to 14%)
Number of cases/rows: variable
Variable List: shear strain in percent (up sweep), shear stress in Pa (up sweep), storage modulus in Pa (up sweep), loss modulus in Pa (up sweep), normal stress in kPa (up sweep), then repeated for the downwards amplitude sweeps
DATA-SPECIFIC INFORMATION FOR: MEMPROTOCOL2 SUMMARIES
Number of variables: 7 (multiplied by the number of read periods, which was 2)
Number of cases/rows: 12
Variable List: read cycle 1 strain in percent, read cycle 1 stress in Pa, read cycle 1 storage modulus in Pa, read cycle 1 loss modulus in Pa, read cycle 1 normal stress in kPa, read cycle 1 shear differential modulus in Pa, read cycle 1 normal differential modulus in Pa, then repeated for the second reading period
