Abstract
Reverse inference is a crucial inferential strategy used in cognitive neuroscience to derive conclusions about the engagement of cognitive processes from patterns of brain activation. While widely employed in experimental studies, it is now viewed with increasing scepticism within the neuroscience community. One problem with reverse inference is that it is logically invalid, being an instance of abduction in Peirce’s sense. In this paper, we offer the first systematic analysis of reverse inference as a form of abductive reasoning and highlight some relevant implications for the current debate. We start by formalising an important distinction that has been entirely neglected in the literature, namely the distinction between weak (strategic) and strong (justificatory) reverse inference. Then, we rely on case studies from recent neuroscientific research to systematically discuss the role and limits of both strong and weak reverse inference; in particular, we offer the first exploration of weak reverse inference as a discovery strategy within cognitive neuroscience.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
When quoting from Peirce’s Collected Papers (CP, Hartshorne et al.1931–1958), we follow the convention of citing the number of the volume followed by the number of the relevant paragraph.
To mention but a few examples of the first tendency, the entry on “Abduction” in the Stanford Encyclopaedia of Philosophy by Douven (2017) only focuses on the “modern” (i.e., strong) sense of abduction, IBE, confining the discussion of the “historical” (i.e., weak) sense to a short supplement. On the opposite side, scholars such as Minnemaier (2004), Campos (2011), and Mcauliffe (2015) have argued against the tendency to equate Peircean (weak) abduction and IBE, claiming that only the first concept can be legitimately called “abduction”. As for the second path mentioned above, for instance Lipton argues for a «version of IBE thus includes two filters, one that selects plausible candidates, and a second that selects from among them» (2004, p. 64) as a unified model of weak ad strong abduction. Similarly, Schurz (2017) explicitly equates abduction with IBE (p. 152) but, at the same time, he carefully analyses the strategic function of abduction, concluding that «the justificatory function of abduction is minor» (p. 153). A minority of scholars avoids overlooking the distinction between weak and strong abduction. For instance, Paavola (2004) explicitly distinguishes between what he calls “Hartmanian abduction” (IBE) and “Hansonian abduction” (weak abduction) even if, following Lipton (2004), he then discusses Hartmanian abduction more as a “method of discovery” than as an instrument for justification.
Indeed, this does not mean forward inference is immune to epistemic risks, as observed by Poldrack & Yarkoni (2016, pp. 589–590). In neuroimaging experiments, the subtraction method is generally used to identify which brain regions are activated by specific cognitive function. This consists in using carefully designed experimental conditions that are supposed to differ only with respect to one process of interest. The subtraction method is problematic because it relies on what has been called the “assumption of pure insertion”, which has been subject to intense criticism in neuroscience (see Poldrack & Yarkoni 2016 for discussion).
Contributors to the debate have proposed quite different approaches to the issue of how to improve strong reverse inference. One sees the main problem in the fact that our cognitive ontology, namely our traditional taxonomy of mental functions and tasks, is outdated and intrinsically defective. The low selectivity of many brain regions might improve when cognitive functions are characterized at a higher level of abstraction (Price & Friston, 2005), or in more precise terms (Poldrack & Yarkoni, 2016). This approach, sometimes labelled “cognitive ontology revision” (Anderson 2015), has motivated the emergence of several computational approaches to mental functions taxonomies, such as the Cognitive Atlas (Poldrack et al., 2011), with the aim of systematizing and improving our ontology of mental concepts and tasks. A second approach tends to question the Bayesian reconstruction originally proposed by Poldrack (2006). Machery (2014), for instance, argues that RI should be reformulated in purely “likelihoodist” terms, thus avoiding the tricky issue of assessing the prior probability of the hypotheses under examination. Others have proposed to conditionalize all probabilities in the Bayesian reconstruction of RI on the specific task used in the study (Del Pinal & Nathan, 2013; Hutzler 2014). A third proposal suggests that reverse inference may be improved, and the selectivity issue mitigated, by shifting the focus of the analysis from isolated brain regions to entire networks of regions (Glymour & Hanson, 2016; Klein, 2012). Finally, the use of multivariate neuroimaging techniques, such as multivoxel pattern analysis (MVPA), has been suggested as a fourth strategy to improve reverse inference, in line with the idea that inferences based on “pattern-decoding” can overcome the problems of more “local” ones (Nathan & Dal Pinal, 2017).
In his paper (2006), Poldrack proposes to address this issue by using one of the several databases of neuroimaging results available on the Internet, i.e., BrainMap (www.brainmap.org), which at that time (Sept. 2005) contained data from 3222 experimental comparisons in 749 published papers. Looking at pairs of experimental comparisons and coordinates of activations included in this database, Poldrack manually calculated the probability of the engagement of language function conditional to the activation of the “Broca’s area” (BA 44) using Bayes theorem. He later compared the posterior probability thus obtained (0.65) to the prior probability of language processes being engaged in a task, conventionally fixed at 0.5, and finally calculated the relative Bayes Factor (2.3) as a proxy of the strength of the reverse inference, resulting in a «positive but relatively weak increase of confidence» (p. 62) in the conclusion.
Note that “reverse inference maps” have been recently renamed “association tests” on the web-based NeuroSynth platform (https://neurosynth.org/faq/).
It is known that the automated lexical algorithms NeuroSynth is based on are not able to extract fine-grained information from texts (e.g., distinguishing different types of memory). Similarly, the algorithms extracting the coordinate of brain activations cannot make basic distinctions such as distinguishing between activations and deactivations (but see Yarkoni et al., 2011).
References
Anderson, D. R. (1987). Creativity and the philosophy of C.S. Peirce. Springer.
Anderson, M. L. (2010). Review of neuroeconomics: Decision making and the brain. Journal of Economic Psychology, 31, 151–154.
Anderson, M. (2015). Minding the brain for a new taxonomy of the mind. Philosophy Compass, 10(1), 68–77.
Bourgeois-Gironde, S. (2010). Is neuroeconomics doomed by the reverse inference fallacy? Mind & Society, 9(2), 229–249.
Campos, D. G. (2011). On the distinction between Peirce’s abduction and Lipton’s Inference to the best explanation. Synthese, 180, 419–442.
Cauda, F., Nani, A., Liloia, D., Manuello, J., Premi, E., Duca, S., Fox, P. T., & Costa, T. (2020). Finding specificity in structural brain alterations through Bayesian reverse inference. Human Brain Mapping, 41(15), 4155–4172.
Costa, T., Manuello, J., Ferraro, M., Liloia, D., Nani, A., Fox, P. T., Lancaster, J., & Cauda, F. (2021). BACON: A tool for reverse inference in brain activation and alteration. Human Brain Mapping, 42(11), 3343–3351.
Crupi, V. (2020). Confirmation. The Stanford Encyclopedia of Philosophy (Spring 2020 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/spr2020/entries/confirmation/>.
Del Pinal, G., & Nathan, M. J. (2013). There and up again: On the uses and misuses of neuroimaging in psychology. Cognitive Neuropsychology, 30(4), 233–252.
Douven, I. (2017), Abduction. The Stanford Encyclopedia of Philosophy (Summer 2017 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/sum2017/entries/abduction/>.
Douven, I., & Schupbach, J. N. (2015). probabilistic alternatives to Bayesianism: The case of explanationism. Cognition, 6, 459.
Frankfurt, H. (1958). Peirce’s notion of abduction. Journal of Philosophy, 55, 593–596.
Genon, S., Reid, A., Langner, R., Amunts, K., & Eickhoff, S. B. (2018). How to characterise the function of a brain region. Trends in Cognitive Sciences, 22(4), 350–364.
Genon, S., Reid, A., Li, H., Fan, L., Müller, V. I., Cieslik, E. C., Hoffstaedter, F., Langner, R., Grefkes, C., Laird, A. R., Fox, P. T., Jiang, T., Amunts, K., & Eickhoff, S. B. (2017). The heterogeneity of the left dorsal premotor cortex evidenced by multimodal connectivity-based parcellation and functional characterisation. NeuroImage, 170, 400–411.
Glymour, C., & Hanson, C. (2016). Reverse inference in neuropsychology. The British Journal for the Philosophy of Science, 67(4), 1139–1153.
Greene, J. D., Sommerville, B., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An FMRI investigation of emotional engagement in moral judgment. Science, 293(5537), 2105–2108.
Hanson, N. R. (1958). Patterns of discovery. Cambridge University Press.
Harman, G. H. (1965). The inference to the best explanation. Philosophical Review, 74, 88–95.
Henson, R. (2006). Forward inference using functional neuroimaging: dissociations versus associations. Trends in Cognitive Sciences, 10(2), 64–69.
Hintikka, J. (1998). What is abduction? The fundamental problem of contemporary epistemology. Transactions of the Charles S. Peirce Society, 34(3), 503.
Hutzler, F. (2014). Reverse inference is not a fallacy per se: Cognitive processes can be inferred from functional imaging data. NeuroImage, 84, 1061–1069.
Klein, C. (2012). Cognitive ontology and region- versus network-oriented analyses. Philosophy of Science, 79(5), 952–960.
Laird, A. R., Lancaster, J. L., & Fox, P. T. (2005). BrainMap: The social evolution of a functional neuroimaging database. Neuroinformatics, 3, 65–78.
Lieberman, M. (2015). Comparing Pain, Cognitive, and Salience Accounts of dACC. A reply to Tal Yarkoni's blog on our paper. URL: https://www.psychologytoday.com/us/blog/social-brain-social-mind/201512/comparing-pain-cognitive-and-salience-accounts-dacc.
Lieberman, M. (2016). More Evidence for a Pain-Related Description of dACC. Selective voxels in the dACC tend to be selective for pain. https://www.psychologytoday.com/ca/blog/social-brain-social-mind/201601/more-evidence-pain-related-description-dacc.
Lieberman, M.D., Burns, S.M., Torre, J.B., Eisenberger, N.I. (2016). Reply to Wager et al.: Pain and the dACC: The importance of hit rate-adjusted effects and posterior probabilities with fair priors. Proc Natl Acad Sci USA, 113(18), E2476–9.
Lieberman, M. D., & Eisenberger, N. I. (2015). The dACC is selective for pain. Proceedings of the National Academy of Sciences, 112(49), 15250–15255.
Lieberman, M. D., Straccia, M. A., Meyer, M. L., Du, M., & Tan, K. M. (2019). Social, self, (situational), and affective processes in medial prefrontal cortex (MPFC): Causal, multivariate, and reverse inference evidence. Neuroscience and Biobehavioral Reviews, 99, 311–328.
Lipton, P. (2004). Inference to the Best Explanation (2nd ed.). Routledge/Taylor and Francis Group.
Machery, E. (2012). Dissociations in neuropsychology and cognitive neuroscience. Philosophy of Science, 79(4), 490–518.
Machery, E. (2014). In defense of reverse inference. The British Journal for the Philosophy of Science, 65(2), 251–267.
Magnani, L. (2001). Abduction. Springer.
Magnani, L. (2009). Creative Abduction and Hypothesis Withdrawal. In J. Meheus & T. Nickles (Eds.), Models of Discovery and Creativity. Origins: Studies in the Sources of Scientific Creativity. Springer.
Mcauliffe, W. (2015). How did Abduction Get Confused with Inference to the Best Explanation? Transactions of the Charles S. Peirce Society, 51(3), 300–319.
McKaugan, D. (2008). From ugly duckling to swan C.S. Peirce, abduction, and the pursuit of scientific theories. Transactions of the Charles S. Peirce Society, 44(3), 446–468.
Minnameier, G. (2004). Peirce-suit of truth—why inference to the best explanation and abduction ought not to be confused. Erkenntnis, 60, 75–105.
Nathan, M. J., & Del Pinal, G. (2017). The future of cognitive neuroscience? reverse inference in focus. Philosophy Compass, 12(7), e12427.
Niiniluoto, I. (1999). Defending abduction. Philosophy of Science, 66(3), 451.
Niiniluoto, I. (2011). Abduction, tomography, and other inverse problems. Studies in History and Philosophy of Science Part A, 42(1), 135–139.
Niiniluoto, I. (2018). Truth-Seeking by Abduction. Springer.
Norton, J. D. (2016). Inference to the best explanation: The general account. In J. Norton (Ed.), The material theory of induction. University of Calgary Press.
Paavola, S. (2004). Abduction as a logic and methodology of discovery: The importance of strategies. Foundations of Science, 9, 267–283.
Paavola, S. (2006). Hansonian and harmanian abduction as models of discovery. International Studies in the Philosophy of Science, 20(1), 93–108.
Pauli, W. M., O’Reilly, R. C., Yarkoni, T., & Wager, T. D. (2016). Regional specialisation within the human striatum for diverse psychological functions. Proceedings of the National Academy of Sciences of the United States of America, 113(7), 1907–1912.
Poldrack, R. (2006). Can cognitive processes be inferred from neuroimaging data? Trends in Cognitive Sciences, 10(2), 59–63.
Poldrack, R. (2008). The role of FMRI in cognitive neuroscience: Where Do We Stand? Current Opinion in Neurobiology, Cognitive Neuroscience, 18(2), 223–227.
Poldrack, R. (2011). Inferring mental states from neuroimaging data: From reverse inference to large-scale decoding. Neuron, 72(5), 692–697.
Poldrack, R., Kittur, A., Kalar, D., Miller, E., Seppa, C., Gil, Y., Parker, D. S., Sabb, F. W., & Bilder, R. M. (2011). The cognitive atlas: Toward a knowledge foundation for cognitive neuroscience. Frontiers in Neuroinformatics, 5, 17.
Poldrack, R., & Yarkoni, T. (2016). From brain maps to cognitive ontologies: Informatics and the search for mental structure. Annual Review of Psychology, 67, 587–612.
Price, C. J., & Friston, K. J. (2005). Functional ontologies for cognition: The systematic definition of structure and function. Cognitive Neuropsychology, 22(3–4), 262–275.
Roche, W., & Sober, E. (2013). Explanatoriness Is evidentially irrelevant, or inference to the best explanation meets Bayesian confirmation theory. Analysis, 73(4), 659–668.
Salmon, W. (2001). Explanation and confirmation: A Bayesian critique of inference to the best explanation. In G. Hon & S. S. Rakover (Eds.), Explanation: Theoretical approaches and applications. Kluwer.
Schackman, A. (2015). The importance of respecting variation in cingulate anatomy: Comment on Lieberman & Eisenberger 2015 and Yarkoni. URL: https://shackmanlab.org/the-importance-of-respecting-variation-in-cingulate-anatomy-comment-on-lieberman-eisenberger-2015-and-yarkoni/
Schickore, J. (2018). Scientific Discovery. The Stanford Encyclopedia of Philosophy (Summer 2018 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/sum2018/entries/scientific-discovery/>.
Schupbach, J. N. (2014). Is the bad lot objection just misguided? Erkenntnis, 79(1), 55–64.
Schupbach, J. N., & Sprenger, J. (2011). The logic of explanatory power. Philosophy of Science, 78(1), 105–127.
Schurz, G. (2008). Patterns of abduction. Synthese, 164, 201–234.
Schurz, G. (2017). Patterns of Abductive Inference. In L. Magnani & L. Bortolotti (Eds.), Springer Handbook of Model-Based Science. Springer.
Sprenger, J., & Hartmann, S. (2019). Bayesian Philosophy of Science. Oxford University Press.
Van Frassen, B. (1989). Laws and Symmetry. Oxford University Press.
Wager, T. (2015). Pain in the ACC. A commentary on Lieberman and Eisenberger. URL: https://www.painresearchforum.org/news/blog/61907-pain-acc
Wager, T. D., Atlas, L. Y., Botvinick, M. M., Chang, L. J., Coghill, R. C., Davis, K. D., Iannetti, G. D., Poldrack, R. A., Shackman, A. J., & Yarkoni, T. (2016). Pain in the ACC? Proceedings of the National Academy of Sciences, 113(18), E2474–E2475.
Weiskopf, D. (2020). Data Mining the Brain to Decode the Mind. In F. Calzavarini, M. Viola, Neural Mechanisms. New Challenges in the Philosophy of Neuroscience. Springer.
Xenophontos, A., Seidlitz, J., Liu, S., Clasen, L. S., Blumenthal, J. D., Giedd, J. N., Alexander-Bloch, A., & Raznahan, A. (2020). Altered sex chromosome dosage induces coordinated shifts in cortical anatomy and anatomical covariance. Cerebral Cortex, 30(4), 2215–2228.
Yarkoni, T. (2015a). No, the dorsal anterior cingulate is not selective for pain: comment on Lieberman and Eisenberger. URL: https://www.talyarkoni.org/blog/2015a/12/05/no-the-dorsal-anterior-cingulate-is-not-selective-for-pain-comment-on-lieberman-and-eisenberger-2015a/
Yarkoni, T. (2015b). Still not selective: comment on comment on comment on Lieberman & Eisenberger. URL: https://www.talyarkoni.org/blog/2015b/12/14/still-not-selective-comment-on-comment-on-comment-on-lieberman-eisenberger-2015b/
Yarkoni, T., Poldrack, R., Nichols, T. E., Van Essen, D. C., & Wager, T. D. (2011). Large-scale automated synthesis of human functional neuroimaging data. Nature Methods, 8(8), 665–670.
Acknowledgements
Previous versions of this article have been presented at various conferences, such as the workshop on "Scientific Errors" (Castelveccana, Italy, 2021) and the symposium on "Reverse Inference: Philosophical and Neuroscientific Perspectives" (ESPP, Online, 2021). We thank the participants at these conferences for their thoughtful comments. We are particularly grateful to Luca Cecchetti, Davide Coraci, Enzo Crupi, Enzo Fano, Diego Marconi, Jan Sprenger, Marco Viola, and two anonymous reviewers for Synthese for very useful comments on this article and/or discussions on its contents.
Funding
Gustavo Cevolani acknowledges financial support from the Italian Ministry of Education, Universities and Research (MIUR) through the grant n. 201743F9YE (PRIN 2017 project ``From models to decisions'').
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Calzavarini, F., Cevolani, G. Abductive reasoning in cognitive neuroscience: weak and strong reverse inference. Synthese 200, 70 (2022). https://doi.org/10.1007/s11229-022-03585-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11229-022-03585-2