The ascending arousal network (AAN) is a subcortical neural network that is critical to consciousness. AAN neurons connect the brainstem to the thalamus, hypothalamus, basal forebrain and cortex, activating cortically-based awareness networks. The reticular core of the AAN was first described by Moruzzi and Magoun in 1949, who coined the classical term "ascending reticular activating system" (Electroencephalogr Clin Neurophysiol 1949;1:455-73). Here, we use the term AAN because many brainstem nuclei that contribute to arousal are located outside of the pontine and midbrain reticular core (e.g., locus coeruleus, parabrachial complex, etc.), and because we believe that the word "network" appropriately connotes the physiological mechanisms by which multiple modular circuits interrelate to enable the emergent property of arousal, and hence consciousness. To date, the majority of studies investigating AAN connectivity have utilized animal models. As a result, current knowledge about the structural and functional connectivity of the human AAN is limited and largely based upon extrapolations from animal data.

We created this AAN atlas to facilitate research into the structural and functional connectivity of the human AAN. The study of AAN "connectomics" has the potential to increase knowledge about arousal physiology in the human brain, as well as arousal pathology in neurological diseases, such as coma and other disorders of consciousness. In addition, the study of AAN connectomics may advance knowledge about reciprocal connectivity between this subcortical arousal network and cortically based awareness networks, such as the default mode network.

In 2012, we released the Harvard AAN Atlas Version 1.0. The atlas was generated using several sources of data: 1) histologic data from a dissected ex vivo human brainstem/diencephalon specimen (Edlow et al. JNEN 2012; https://pubmed.ncbi.nlm.nih.gov/22592840/); 2) correlative diffusion data from the same specimen (scanned prior to serial sectioning and staining); and 3) cross-reference to the Paxinos human brainstem atlas (Paxinos G, Xu-Feng H, Sengul G, Watson C. Organization of Brainstem Nuclei, in The Human Nervous System, 3rd ed. Mai JK and Paxinos G eds. Amsterdam: Elsevier, 2011). Additional details about the development of Version 1.0 are provided at https://www.nmr.mgh.harvard.edu/resources/aan-atlas.

Here, we release Version 2.0 of the atlas, which incorporates updates to AAN node neuroanatomy and nomenclature. Neuroanatomic updates are based on new immunostaining data from human brainstem specimens, which can be found at https://histopath.nmr.mgh.harvard.edu. Nomenclature updates are to ensure consistency with the Paxinos 2011 Atlas (Paxinos G, Xu-Feng H, Sengul G, Watson C. Organization of Brainstem Nuclei, in The Human Nervous System, 3rd ed. Mai JK and Paxinos G eds. Amsterdam: Elsevier, 2011).

The updates in Version 2.0 are:

1) Updates to Node Nomenclature

• mRt: cuneiform/subcuneiform nucleus (CSC) --> mesencephalic reticular formation (mRt)
• PTg: pedunculopontine nucleus (PPN) --> pedunculotegmental nucleus (PTg)

2) Updates to Node Neuroanatomy

• mRt: In Version 1.0, the CSC was localized to the rostral midbrain, based on Moruzzi and Magoun’s seminal study in which stimulation of the rostral midbrain tegmentum led to arousal and cortical activation in lightly anesthetized cats.  In Version 2.0, we renamed this node mRt and changed its anatomic boundaries such that it now extends throughout the entire rostro-caudal axis of the midbrain tegmentum, consistent with the Paxinos Atlas.  Specifically, our mRt node now contains all of the following subnuclei described by Paxinos 2011: CnF, isRt, PrCnF, mRt, CeMe, and p1Rt.

• VTA: The VTA node in Version 2.0 has two major differences compared to the VTA ROI used in Version 1.0: 1) its lateral extent and 2) its rostral extent.  In Version 1.0, we did not include the lateral wing-like extensions of the parabrachial pigmented nucleus of the VTA (PBP), as shown in Paxinos, because this subregion is difficult to differentiate from the nearby substantia nigra.  However, with our new tyrosine hydroxylase immunostaining, and with reference to the Paxinos atlas, we were able to make this distinction between PBP and SN in Version 2.0.

  With respect to the rostral extent of the VTA node, in Version 1.0 we based the neuroanatomic borders of the VTA on hematoxylin-and-eosin-stained sections from human brainstem specimens, which revealed the highest density of catecholamine neurons, as demonstrated by the presence of neuromelanin, in the caudal midbrain.  However, we recognize that VTA neurons have been identified in the rostral mesencephalon by other laboratories using a variety of staining techniques, such as tyrosine hydroxylase.  Thus, we performed new tyrosine hydroxylase stains on midbrain sections from human brainstem specimens.  These stains revealed a distribution of VTA neurons that was consistent with that described in prior studies by Oades, Halliday, Pearson, and colleagues.  Therefore, we extended the VTA node in Version 2.0 such that it runs throughout the entire rostro-caudal axis of the midbrain.  We do not include the rostral linear nucleus as part of the VTA, although some dopaminergic cells may reside in this region, because this nucleus is primarily serotonergic.

  Finally, we edited the midline portion of the VTA node, based on the observation that VTA neurons in the midline of the midbrain were only seen at the level of the red nucleus, not at the level of the superior cerebellar peduncle.  This immunostaining observation is consistent with the VTA anatomic borders in the Paxinos atlas along the rostro-caudal axis of the midbrain.  Thus, we removed the midline VTA voxels at the level of the superior cerebellar peduncle in Version 2.0.

  In summary, the VTA node in Version 2.0 contains the following nuclear subregions described by Paxinos: parainterfascicular nucleus (PIF); paranigral nucleus of ventral tegmental area (PN); ventral tegmental area (VTA); ventral tegmental area, rostral part (VTAR); and parabrachial pigmented nucleus of the VTA (PBP).  

• PTg: Deleted one voxel in the PTg at MNI axial level z53 and reassigned this voxel to the VTA, based on tyrosine hydroxylase staining data and based on the anatomic boundaries in the Paxinos atlas.
• LDTg: Whereas LDTg was not included in Version 1.0, we have added LDTg to Version 2.0.