Manuscript template
Prepare the main manuscript and keep author identities in the separate title page.
Download manuscript templateFor authors
A practical route through scope, manuscript files, scientific reporting and the declarations needed for a fair review.
Prepare an accurate AI-use declaration, verify references and explain the provenance of figures, data and code.
Recommended before submission. Check sections, declarations, citations and references, then prepare a Word draft with this journal’s template.
The journal is led by Editor-in-Chief Dr Rekha C. R. Manuscript reception remains closed while the submission workflow is prepared. You may use the guidance and Manuscript Pre-check to prepare your work; preparation is not a journal submission. The opening date and authorised submission route will be announced here.
Downloading a template or preparing evidence does not submit a manuscript. The formal submission route will be announced when the journal opens.
Use these editable files alongside the instructions below. They are preparation resources; they do not create an active submission.
Prepare the main manuscript and keep author identities in the separate title page.
Download manuscript templateRecord the author list, affiliations, contributions and relevant declarations separately.
Download title pageExplain the contribution, journal fit and related work for editorial assessment.
Download cover letterAddress each review point and identify revisions or explain a reasoned disagreement.
Download response to reviewersPrepare an anonymised manuscript, a separate title page, a cover letter and any supporting files needed to assess the work. The title page should give the full journal title, manuscript title, article type, author names and affiliations, corresponding author details, contribution statement and relevant declarations. Keep identities and identifying acknowledgements out of the review manuscript. Anonymise document properties and explain any repository link that could reveal authorship. Ethics, data-access and methodological declarations needed to assess the science must remain in anonymised form in the review manuscript. Provide identifying approvals or other confidential evidence separately through the authorised editorial route.
Prepare an editable DOCX manuscript using the journal structure; a review PDF may accompany it. Authors using LaTeX or another specialist format should seek an agreed alternative route before submission and provide a compiled review PDF and source when requested. Supply original figure files and an accurate reference list for production. Initial submissions may use a consistent readable reference style. The final house style is specified during production. References to data and software should identify the version or release and a persistent identifier where available.
Use a specific title, an informative abstract and descriptive keywords. State the research question, its relationship to prior work, the methods and assumptions, the evidence, limitations and conclusions. For empirical work, distinguish results from interpretation. For theoretical, conceptual or qualitative work, use a structure that makes the reasoning and evidence traceable. Include all applicable declarations and cite supplementary material in the manuscript.
There is no fixed initial-submission word limit in this guide. Provide a word count and use the length needed for a complete, focused account. The editor may request shortening or restructuring. A later published article-type limit will apply to submissions received after its stated effective date; any exceptional change for a manuscript already under consideration must be communicated to its authors.
A preparation file or saved evidence record does not constitute a journal submission. Formal submission requires the journal's announced submission route, author approval and completion of the required declarations. Authors should use the manuscript reference in subsequent correspondence and identify the version to which a query relates.
A revised submission should include an updated anonymised manuscript, any revised supporting files and a response addressing every editorial and reviewer point. Identify where changes were made and explain any reasoned disagreement. Keep the response professional and provide evidence for disputed technical issues. Do not silently substitute datasets, analyses or author lists; disclose consequential changes for editorial assessment.
Journal-specific standards
Use the guidance relevant to your design. A formal, conceptual or behavioural study should not add unrelated experiments merely to resemble another article type.
Begin with a clear problem and contribution. The abstract should identify the research question, approach, principal finding and its practical or theoretical limits. An introduction should explain the gap using relevant literature, rather than rely on unsupported claims that a method is novel or universally superior.
A suitable structure includes related work, problem formulation, methods, evaluation or analytical results, discussion and limitations. Adapt the structure to the study design. Define notation, explain criterion direction and units, and keep indispensable assumptions and evidence in the main narrative. Appendices can hold derivations and implementation details but should not conceal a condition on which the main conclusion depends.
Describe the sampling unit, task and intended generalisation setting. State how training, validation and test partitions were created, including grouping by person, institution, device or time where dependence makes ordinary random splitting inappropriate. Fit preprocessing and feature selection using the permitted training information. Explain steps taken to prevent leakage through duplicates, labels, overlapping records, repeated participants or benchmark contamination.
Choose baselines that address the same question and explain exclusions. Report tuning procedures and resource budgets sufficiently to assess whether comparisons received comparable treatment. Identify random seeds, repeated runs, stopping conditions and the evaluation set used for model selection. For language and generative models, record provider, model identifier, accessible version or date, prompts, decoding settings, tool access and output handling where relevant.
Use metrics that reflect the task and possible consequences of errors. Explain averaging, thresholds and treatment of missing, invalid or abstained outputs. Report distributions, intervals or other uncertainty estimates appropriate to the design. Statistical tests should respect dependence and repeated comparison. An improvement in a single average score does not establish superiority across populations, tasks or operating conditions.
Define alternatives, criteria, decision context and the intended meaning of the output. Explain how criteria were identified and whether any overlap or dependence affects interpretation. State measurement scales, benefit or cost direction, normalisation, missing-value handling and any conversion of linguistic or fuzzy judgements. Document the mathematical steps and assumptions connecting inputs to weights, scores, ranks or recommendations.
When expert judgement is used, explain expertise requirements, recruitment, eligibility, number of participants and the elicitation procedure. Describe instructions, anchors, rounds, aggregation and disagreement handling. Distinguish real participants from simulated judgements. Simulated experts can support a clearly labelled methodological demonstration but cannot establish actual expert consensus or substitute for an empirical elicitation study.
Examine sensitivity to plausible changes in judgements, weights, normalisation choices and uncertain inputs. Explain ties, rank reversals, conflicting criteria and stability where relevant. Compare methods on a reasoned basis and discuss what agreement or disagreement means. A numerical ranking should not be described as objective truth when it depends on preferences, assumptions or an incomplete set of alternatives.
State the domain of every formal claim and distinguish proved results from conjectures, approximations and experimental observations. Explain the assumptions under which existence, convergence, consistency or optimality follows. When an algorithm is heuristic, describe what is and is not guaranteed. Negative results and counterexamples are valuable when their conditions and implications are clear.
Discuss uncertainty at its relevant sources, including sampling, model estimation, preference elicitation, distribution shift and measurement. Explain whether reported confidence, credibility or stability measures support the interpretation given to them. Decision-makers should be able to distinguish uncertainty about evidence from disagreement about values.
Provide an environment description identifying software, dependencies, hardware and computational requirements. Include instructions connecting data preparation, training or optimisation, evaluation and reported tables or figures. Document checkpoints, configuration files and versions. Where commercial APIs or changing online services are used, explain which conditions can be reproduced and what may have changed.
For datasets and trained models, identify origin, collection or training information available to the authors, access restrictions and applicable licences. Explain permission for the uses undertaken and any uncertainty about third-party training material. Distinguish real, synthetic, augmented and AI-generated data. Restricted materials require a meaningful access explanation or safe supporting evidence, rather than a claim that openness overrides participant rights or confidentiality.
Where the work creates a meaningful risk of harm or misuse, provide a proportionate account of affected groups, plausible failure conditions and mitigation. Relevant issues may include discrimination, unsafe automation, intrusive inference, surveillance or manipulation of decisions. Explain the connection to the actual research; avoid speculative claims about remote applications unsupported by the study.
A model evaluated retrospectively or in a controlled experiment should not be presented as ready for unrestricted deployment. Explain the additional evidence needed for the proposed use, including human oversight or domain validation where appropriate. Editors may request specialist assessment when ordinary methodological review cannot resolve a material ethical concern.
Adapt these examples to what actually occurred. Keep science-relevant information in the anonymised manuscript and provide identifying evidence separately when needed.
The materials supporting [specified results] are available at [repository and persistent identifier], version [version]. The repository contains [data or access instructions], [code], [configuration] and [environment details]. Access to [restricted material] is limited because [reason]; requests may be directed through [authorised access route].
Judgements were obtained from [number] participants recruited using [eligibility and recruitment method]. Elicitation used [instrument and procedure]. Participation, consent and ethics review or exemption were addressed through [applicable details]. [Aggregation method] combined the responses, and [analysis] examined disagreement and sensitivity.
The judgements in [specified analysis] are simulated inputs constructed for a methodological demonstration. They are not observations from recruited experts and do not establish expert consensus. Input construction and the tested sensitivity conditions are described in [section or resource].
Publishing information
Read the journal’s publication frequency, charges, editorial policies and discovery information.