Functionally dissociating ventro-dorsal components within the rostro-caudal hierarchical organization of the human prefrontal cortex
Introduction
Within the lateral prefrontal cortex (PFC), different levels of cognitive control are assumed to be hierarchically organized along a rostro-caudal axis, with rostral parts of the PFC performing highly abstract levels of behavioral control and caudal parts carrying out concrete action selection in a temporally confined context (Badre et al., 2009; Blumenfeld et al., 2013; Koechlin et al., 2003, 1999; Voytek et al., 2015). Evidence for this hierarchical organization of neural processing has been provided by task-based functional magnetic resonance imaging (fMRI) studies (Badre & D'Esposito, 2007; Bahlmann et al., 2015; Koechlin et al., 1999, for a review see Badre & D'Esposito, 2009) but is also supported by lesion data (Azuar et al., 2014) and transcranial magnetic stimulation (Nee and D'Esposito, 2016). However, other studies showed that rostral PFC regions can also be recruited by concrete action selection (Crittenden and Duncan, 2014) and that the temporal, rather than the spatial activation profile of specific PFC regions is modulated by maintenance demands, irrespective of the level of abstraction (Reynolds et al., 2012). Tracer studies in monkeys further demonstrated that the structural network in the PFC does not follow a strict rostro-caudal organization (Goulas et al., 2014). The extent to which the PFC is organized along a rostro-caudal axis hence constitutes a matter of debate.
Beyond functional gradients along a rostro-caudal axis, the structural and functional organization of the PFC has also been subject to anatomically detailed characterizations along a ventro-dorsal axis (see Tanji and Hoshi, 2008 and Petrides, 2005 for reviews). In this respect, it has been demonstrated that potentially separable rostro-caudal streams of processing are present in the ventral and dorsal convexity of the lateral PFC (Blumenfeld et al., 2018, 2013). Using a resting-state fMRI paradigm, Blumenfeld et al. (2013) found parallel ventral and dorsal networks that were interconnected in caudal but not in rostral PFC regions. Bahlmann et al. (2015) further suggested that rostro-caudally organized functional networks in ventral and dorsal PFC adapt their ventro-dorsal segregation dynamically to be operative on the highest level of the rostro-caudal axis that is currently engaged in the task, whereas on lower levels processing is integrated across ventral and dorsal areas. The lateral PFC thus seems to comprise parallel rostro-caudal pathways which appear anatomically separable along a ventro-dorsal axis but functionally interact to subserve goal-directed behavior. While this functional interaction has been proposed to be orchestrated by the rostral-most part of the lateral PFC (e.g. Ramnani and Owen, 2004; Wendelken et al., 2012; for a recent review on the function of the frontopolar cortex see Mansouri et al., 2017), recent evidence suggests that the apex of the prefrontal hierarchy actually resides in the mid-lateral rather than the rostral PFC (Margulies et al., 2016; Nee and D'Esposito, 2016; for a review see Badre and Nee, 2018).
Taken together, an abundance of fMRI studies demonstrate the gradual activation along the rostro-caudal and the ventro-dorsal axes of the PFC by task-related factorial designs (e.g. Bahlmann et al., 2015) as well as the functional connectivity between the respective regions by correlation analyses of resting-state activity (e.g. Taren et al., 2011). These studies argue for a hierarchical functional organization of the PFC. However, to fully understand the mechanisms and functional pathways that subserve cognitive functions requires to complement these correlation- and activation-based analyses by the inference of the actual direction of influences and the demonstration of the implied propagation of neural activity along a rostro-to-caudal gradient of hierarchical control within the lateral PFC. While the slow hemodynamic response is well captured by sampling intervals between 0.5 and 2 Hz as provided by conventional fMRI (Logothetis, 2008), functional interactions between brain regions appear on a much smaller temporal scale (Stokes et al., 2013). Reliably inferring directed functional connections from such very short delays between neural activity in different regions requires much faster sampling of at least 10 Hz (Mader et al., 2008; Roebroeck et al., 2005). Simplifying the problem of inferring directionality down to the detection of short delays between oscillations (Granger, 1969), the need for a sufficiently high temporal resolution can be easily illustrated by plotting two noisy sine waves with a small phase shift using different sampling rates. A phase shift which is entirely obscure when sampled at .5 Hz can become highly apparent when sampled at 10 Hz (Supplementary Fig. S1).
Similar to fMRI, functional near-infrared spectroscopy (fNIRS) relies on the neuro-vascular coupling and measures the hemodynamic response but in contrast to fMRI it is based on the differential absorption properties of oxygenated and deoxygenated hemoglobin (Scholkmann et al., 2014; Strangman et al., 2002). Multiple light sources and detectors transcranially measure absorption changes elicited by changes in cortical oxygenation at sample frequencies up to 250 Hz (Scholkmann et al., 2014). Thus, fNIRS overcomes the limited temporal resolution of fMRI and provides sufficiently high spatial resolution (here 2.1 cm) to allow for inference of directed functional interactions along rostro-caudal and ventro-dorsal axes in the PFC.
Here we used fast optical imaging with multi-channel fNIRS and measures of directed coherence (DC) (Schelter et al., 2006) to estimate the propagation of neural activity across the PFC and to provide complementary evidence for the predicted influences within and between parallel rostro-caudal signaling pathways in the ventral and dorsal PFC (Bahlmann et al., 2015; Blumenfeld et al., 2013; Bunge et al., 2005; Wendelken et al., 2012). We expected to reveal (i) a predominant rostral to caudal direction of influences within the PFC and (ii) a separation into a ventral and a dorsal component.
Section snippets
Experimental design
Subjects between 19 and 26 years of age were recruited from the University of Freiburg provided that they were German native speakers and fulfilled MRI safety criteria. Exclusion criteria concerned current or previous psychiatric/neurological disease, use of psychotropic medication, and color blindness. Thirty-one subjects participated in two 24-min fNIRS measurements (one week apart) and additionally underwent MRI and neuropsychological assessments that were conducted as a part of a larger
Results
Due to the high number of observations all fixed effects terms in both models were significant (p < .05); we therefore only report significant digits (Clymo, 2014) of least square means of DC values (DCLSM) ± standard errors and post-hoc tests of interest in the text and refer the reader to Supplementary Table S1 for a detailed overview of effect statistics. In the following, ΔDCLSM denotes contrasts (pairwise comparisons) of DCLSM values and ΔΔDCLSM denotes interaction contrasts (pairwise
Discussion
Taking advantage of the sufficiently high temporal and spatial resolution of multi-channel fNIRS, the present study used directed coherence as a measure of influences between brain regions to assess the functional networks of the PFC (Medvedev, 2014). Showing that activity in caudal PFC is modulated by activity in its more rostral parts, the present data provide complementary evidence for the intrinsic rostro-caudal functional hierarchy within the PFC, as predicted by extant models of
Conclusion
Using the methodological framework of fast sampling multi-channel fNIRS and a frequency-domain measure of directed functional connectivity, we provide explicit evidence for a rostral-to-caudal processing hierarchy in the PFC. Consistent with extant models of prefrontal organization, this hierarchy is dissociated into a ventral and a dorsal component and peaks in the mid-dorsolateral PFC, which exerts the highest level of cognitive control (Badre and Nee, 2018).
Conflicts of interest
The authors declare no competing financial interests.
Acknowledgement
This work was supported by a grant of the BrainLinks-BrainTools Cluster of Excellence funded by the German Research Foundation (DFG, grant number EXC 1086).
References (73)
A fast diffeomorphic image registration algorithm
Neuroimage
(2007)- et al.
Testing the model of caudo-rostral organization of cognitive control in the human with frontal lesions
Neuroimage
(2014) - et al.
Frontal cortex and the hierarchical control of behavior
Trends Cognit. Sci.
(2018) - et al.
Dissociable controlled retrieval and generalized selection mechanisms in ventrolateral prefrontal cortex
Neuron
(2005) - et al.
Brain dynamics during natural viewing conditions--a new guide for mapping connectivity in vivo
Neuroimage
(2005) - et al.
The effects of lateral prefrontal transcranial magnetic stimulation on item memory encoding
Neuropsychologia
(2014) - et al.
Prefrontal and hippocampal contributions to visual associative recognition: interactions between cognitive control and episodic retrieval
Brain Cognit.
(2004) - et al.
The gateway hypothesis of rostral prefrontal cortex (area 10) function
Trends Cognit. Sci.
(2007) - et al.
Rostrolateral prefrontal cortex involvement in relational integration during reasoning
Neuroimage
(2001) - et al.
Functional near infrared spectroscopy (NIRS) signal improvement based on negative correlation between oxygenated and deoxygenated hemoglobin dynamics
Neuroimage
(2010)
Maintenance versus manipulation of information held in working memory: an event-related fMRI study
Brain Cognit.
The effect of filtering on Granger causality based multivariate causality measures
Neuroimage
Distinct functional connectivity associated with lateral versus medial rostral prefrontal cortex: a meta-analysis
Neuroimage
A functional gradient in the rodent prefrontal cortex supports behavioral inhibition
Curr. Biol.
An information theoretical approach to prefrontal executive function
Trends Cognit. Sci.
Correlations in low-frequency BOLD fluctuations reflect cortico-cortical connections
Neuroimage
Does the resting state connectivity have hemispheric asymmetry? A near-infrared spectroscopy study
Neuroimage
Mapping directed influence over the brain using Granger causality and fMRI
Neuroimage
Testing for directed influences among neural signals using partial directed coherence
J. Neurosci. Methods
A review on continuous wave functional near-infrared spectroscopy and imaging instrumentation and methodology
Neuroimage
Probing the reproducibility of quantitative estimates of structural connectivity derived from global tractography
Neuroimage
Dynamic coding for cognitive control in prefrontal cortex
Neuron
Non-invasive neuroimaging using near-infrared light
Biol. Psychiatry
Dissociation and convergence of the dorsal and ventral visual working memory streams in the human prefrontal cortex
Neuroimage
Is the rostro-caudal axis of the frontal lobe hierarchical?
Nat. Rev. Neurosci.
Functional magnetic resonance imaging evidence for a hierarchical organization of the prefrontal cortex
J. Cognit. Neurosci.
Mechanisms of hierarchical reinforcement learning in cortico-striatal circuits 2: evidence from fMRI
Cerebr. Cortex
Hierarchical cognitive control deficits following damage to the human frontal lobe
Nat. Neurosci.
The rostro-caudal Axis of frontal cortex is sensitive to the domain of stimulus information
Cerebr. Cortex
Levels of integration in cognitive control and sequence processing in the prefrontal cortex
PLoS One
Fitting linear mixed-effects models using lme4
J. Stat. Software
Manual for the Beck Depression Inventory–ii
Functional connectivity in the motor cortex of resting human brain using echo-planar MRI
Magn. Reson. Med.
Quantitative anatomical evidence for a dorsoventral and rostrocaudal segregation within the nonhuman primate frontal cortex
J. Cognit. Neurosci.
Lateral prefrontal cortex is organized into parallel dorsal and ventral streams along the rostro-caudal axis
Cerebr. Cortex
Prefrontal cortex and long-term memory encoding: an integrative review of findings from neuropsychology and neuroimaging
Neuroscientist
Cited by (10)
Cortical hierarchy, dual counterstream architecture and the importance of top-down generative networks
2021, NeuroImageCitation Excerpt :The controversial issue of frontal cortex and hierarchical control shows promise of resolution via a revision of the concept of a unitary model ensuring a unidimensional gradient. Recent reports favor distinct networks that are proposed to interact thereby ensuring a global hierarchical structure (Schumacher et al., 2019). Nevertheless, the mismatch between the multistage cascade architecture mediating a temporal organization of cognitive control and inter-areal connectivity contrasts with the situation in the visual cortex where there is a smooth increase in receptive field size ascending the Felleman and Van Essen cortical hierarchy (Roelfsema and de Lange, 2016).
From ATOM to GradiATOM: Cortical gradients support time and space processing as revealed by a meta-analysis of neuroimaging studies
2021, NeuroImageCitation Excerpt :Furthermore, our theory updates the ATOM view, showing that space and time representations are distributed in the brain along an anterior-posterior axis in the SMA regions, and a dorsal-ventral axis in the fronto-parietal regions. Based on the evidence for an anterior-posterior and dorsal-ventral processing hierarchy in the PFC (Badre et al., 2009; Schumacher et al., 2019), it is plausible to assume a hierarchy in space-time processing as well, where temporal inputs are processed and manipulated “in ministerium” of forming/enriching spatial representations to greater extent than vice versa. Gradient organization thus facilitates hierarchical transformations by enabling space- and time-related neural populations to interact with each other over minimal distances (Harvey and Dumoulin, 2017).
Behavioral, cognitive and histological changes following neonatal anoxia: Male and female rats’ differences at adolescent age
2019, International Journal of Developmental NeuroscienceCitation Excerpt :Interestingly, we found reduction in the number of cells in the ventral CA3-2 subfields in anoxic male rats. The ventral CA3-2 is linked to the contextual fear conditioning (Mastrodonato et al., 2018) and emotional responses, such as learned approach-avoidance conflict processing (Schumacher et al., 2018). In this way, we can suggest that neonatal anoxia might lead to disruption in these behavioral parameters.
Frontopolar Cortex Specializes for Manipulation of Structured Information
2022, Frontiers in Systems Neuroscience