Skip to main content

Morphological Features of Human Dendritic Spines

  • Chapter
  • First Online:
Dendritic Spines

Part of the book series: Advances in Neurobiology ((NEUROBIOL,volume 34))

Abstract

Dendritic spine features in human neurons follow the up-to-date knowledge presented in the previous chapters of this book. Human dendrites are notable for their heterogeneity in branching patterns and spatial distribution. These data relate to circuits and specialized functions. Spines enhance neuronal connectivity, modulate and integrate synaptic inputs, and provide additional plastic functions to microcircuits and large-scale networks. Spines present a continuum of shapes and sizes, whose number and distribution along the dendritic length are diverse in neurons and different areas. Indeed, human neurons vary from aspiny or “relatively aspiny” cells to neurons covered with a high density of intermingled pleomorphic spines on very long dendrites. In this chapter, we discuss the phylogenetic and ontogenetic development of human spines and describe the heterogeneous features of human spiny neurons along the spinal cord, brainstem, cerebellum, thalamus, basal ganglia, amygdala, hippocampal regions, and neocortical areas. Three-dimensional reconstructions of Golgi-impregnated dendritic spines and data from fluorescence microscopy are reviewed with ultrastructural findings to address the complex possibilities for synaptic processing and integration in humans. Pathological changes are also presented, for example, in Alzheimer’s disease and schizophrenia. Basic morphological data can be linked to current techniques, and perspectives in this research field include the characterization of spines in human neurons with specific transcriptome features, molecular classification of cellular diversity, and electrophysiological identification of coexisting subpopulations of cells. These data would enlighten how cellular attributes determine neuron type-specific connectivity and brain wiring for our diverse aptitudes and behavior.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 149.00
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Hardcover Book
USD 199.99
Price excludes VAT (USA)
  • Durable hardcover edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Notes

  1. 1.

    “Clearly, only humans sit around the fire (or dinner table) to tell each other jokes and stories about past glories or future plans, and only a human would eagerly read what Owen wrote about primate brains 140 years ago. Moreover, only humans use general engineering skills to overcome environmental challenges that other animals can solve solely through evolution by natural selection. These are, of course, merely some of the major differences between human” and other animals (Striedter 2004).

  2. 2.

    “In particular, the activity of one spine can modulate the plasticity of neighboring spines through the mutual sharing of plasticity-related proteins or through the activation of synchronized synaptic inputs. These changes occur across different time scales and typically result in the spatial organization of spines into groups or clusters. This phenomenon suggests that spines do not act as single functional units but are part of a complex network that organizes spines in groups to optimize the connectivity patterns between dendrites and surrounding axons” (Mijalkov et al. (2021) and references therein).

  3. 3.

    Neuroanatomical and functional heterogeneity may exist within cortical areas and their subdivisions (e.g., see Vogt 2015) which refers to the surface location, macroscopic aspect, and dimension (e.g., Scheperjans et al. 2008; Bruner et al. 2014, 2017a, b); the corresponding function depending on experience, emotion, and behavior (e.g., in musicians, Gaser and Schlaug 2003; Omigie 2016; Bouhali et al. 2020); and variable gray matter width, heterogeneous cytoarchitectonic width, cellular composition, and neuropil package (Pandya et al. 2015; Mai et al. 2016; Triarhou, 2009 and the von Economo and Koskinas’ atlas of cytoarchitectonics of the adult human cerebral cortex; note a distinctive layer IV in motor cortex discussed by Yamawaki et al. (2014); and for data indicating that “cortical functions should rather focus on circuits specified by functional cell type composition than mere laminar location,” see Guy and Staiger (2017)).

  4. 4.

    In the human and mouse temporal cortex, “the most-divergent gene families include neurotransmitter receptors, ion channels, extracellular matrix elements, and cell-adhesion molecules” (Hodge et al. 2019). There are “species differences in the expression of genes encoding 5-HT1 receptor family subunits. In mouse layer V extratelencephalic-projecting neurons, HTR1A and HTR1F were the dominantly expressed subunits, whereas, in human layer V extratelencephalic-projecting neurons, HTR1E (which is absent in the mouse genome) and HTR1F were highly expressed, with little HTR1A expression. These data suggest that human and rodent layer V extratelencephalic-projecting neurons likely share some similar distinctive intrinsic membrane properties and responses to neuromodulation in comparison to neighboring layer V intratelencephalic-projecting neuron types. In contrast, cross-species gene expression differences among layer V extratelencephalic-projecting neurons in mouse versus human highlight areas of potential phenotypic divergence” (Kalmbach et al. 2021). Therefore, extrapolations from animal models data on neuronal morphological, connectional, and functional features have to be done carefully when assuming similar implications to the human cerebral cortex (Hodge et al. 2019). Moreover, it is apparent “that genome-wide quantitative differences in expression profiles between species must also be considered when assessing the fine-tuned functional properties of a given cell type in different species... even classically defined, cerebellar cell types differ between mouse and human by expression of hundreds of orthologous genes... granule cell or astrocyte gene expression profiles can vary between species, or even in individual cells of a type, without losing their cell type identity. Studies of sixteen human postmortem brains revealed gender-specific transcriptional differences, cell-specific molecular responses to aging, and the induction of a shared, robust response to an unknown external event evident in three donor samples” (Xu et al. 2018).

  5. 5.

    The frontal, parietal, and temporal association cortices are larger in humans relative to those of other primates (Kolb and Whishaw (2021); see also Striedter (2004) for a critical discussion and findings on thalamic nuclei and cerebellar hemispheres, and Rasia-Filho et al. (2021) for a discussion on amygdaloid nuclei development with allocortex and neocortex evolved functions). The precuneus in the posteromedial parietal cortex is another cortical area with marked expansion in our species (Bruner et al. 2017a, b; see also Messina et al. 2023).

  6. 6.

    Furthermore, “dendrites of layer II/III human pyramidal cortical neurons are more excitable, generating multiple dendritic Ca2+ spikes upon current injection. Dendritic APs in apical tuft dendrites were also found to be sharply tuned to specific input strengths... As a result, when inputs exceed optimal input strength, the amplitude of the dendritic Ca2+ AP is reduced. A striking consequence of this change in electrical properties of distal dendrites is that it enables human cortical pyramidal neurons to execute XOR logical operations in apical tuft dendrites, thereby extending the computational repertoire beyond simple AND/OR operations... As such, superficial and deep layer cortical pyramidal neurons may have evolved distinct mechanisms in response to the growing size of the cortex that depends on the computational role they play in the circuit. Each of these mechanisms may have enhanced the computational power in distinct ways. By increasing and tuning dendritic excitability of layer II/III pyramidal neurons, distinct logical operations can be performed on a dendritic level that otherwise would require the implementation of a complex neuronal circuit. In contrast, the isolated nature of distal dendrites in deep layer cortical neurons may provide a distinct compartment for parallel processing of information” (Schmidt and Polleux 2022).

  7. 7.

    Accordingly, “the stronger connections in human are not due to a larger number of synaptic contacts. Rather, it is explained by the larger presynaptic active zones and PSDs in human that may allow higher release probability as well as more neurotransmitter release and binding (Benavides-Piccione et al. 2002; Yakoubi et al. 2019b), ultimately leading to larger synaptic conductance at human synapses” (Hunt et al. 2022). Specifically, “beside similarities, human synaptic boutons, although comparably small (approximately 5 μm), differed substantially in several structural parameters, such as the shape and size of active zones, which were on average 2 to 3-fold larger than in experimental animals. The total pool of synaptic vesicles exceeded that in experimental animals by approximately 2 to 3-fold, in particular the readily releasable and recycling pool by approximately 2 to 5-fold, although these pools seemed to be layer-specifically organized” (Rollenhagen et al. 2020). See additional relevant data comparing synaptic structure in humans and rats in Molnár et al. (2016). Importantly, layer-specific synaptic transmission structural parameter differences in temporal lobe layers IV and V, with larger values in layer IV, are suggestive of different neurotransmission efficacy, strength, and plasticity between human cortical layers (Yakoubi et al. 2019a, b). Human layer IV excitatory synaptic boutons may act as “amplifiers” of signals from the sensory periphery to integrate, synchronize, and modulate intra- and extracortical synaptic activity (Yakoubi et al. 2019b).

  8. 8.

    For more information on human multimodal, higher-order processing and connected large-scale networks for the default mode and resting state, sensorimotor, executive control, subcortical and cortical emotional processing, detection of salience and attention, task-related and (abstract) cognitive activity, basal ganglia circuitry, and social behavioral display, see Goulden et al. (2014), van den Heuvel et al. (2015), Margulies et al. (2016), Ng et al. (2016), Diano et al. (2017), Bagarinao et al. (2019), Leopold et al. (2019), Bagarinao et al. (2020), Deming and Koenigs (2020), Shafiei et al. (2020), Bruton (2021), Hidalgo-Lopez et al. (2021), and Kolb and Whishaw (2021).

  9. 9.

    “In the course of studying changes in gray matter, it has also been possible to distinguish group differences between healthy children and those displaying neurodevelopmental disorders (e.g., Giedd and Rapoport 2010). Fortunately, cortical abnormalities in children with neurodevelopmental disorders may not be permanent. Shaw et al. (2007) followed about 500 children, some typically developing and others with attention-deficit/hyperactivity disorder (ADHD). They found that reduced volume of gray matter in the prefrontal cortex was not permanent but rather reflected a delay in cortical development by about 2 ½ years, suggesting that ADHD is characterized by a delay rather than a deviance in cortical development... The association between delayed cortical development and ADHD is a novel hypothesis that will guide both research and treatment for the foreseeable future” (Kolb and Whishaw 2021). It is also open to determine the role of dendritic spines in this condition.

  10. 10.

    “Topologically direct interconnections between spatially remote brain regions will increase the efficiency of information processing, which is expected to yield benefits in terms of adaptive behavior. Brain networks can therefore be said to negotiate an economical trade-off between minimizing physical connection cost and maximizing topological value” (Bullmore and Sporns 2012).

  11. 11.

    There are open questions about the density and types of dendritic spines - if they indeed exist - in the morphologically heterogeneous populations of enteric neurons in our species. While submucosal neurons can be nondendritic, myenteric cells display around 20 very short dendrites arising from the cell body. These neurons were named Dogiel “spiny” (or with a “thorny” aspect) type I neurons. Considering the axonal and dendritic shapes, other Dogiel types I neurons were classified as “stubby,” with short and partly stubby or partly lamellar dendrites, or “hairy” cells, with short and very thin dendrites. Dogiel type II neurons can show up to 16 dendrites, but these processes “look like” and behave as axons conducting APs. Types III and IV neurons are long dendritic, uniaxonal neurons. Type V neurons appear as unipolar cells with a single stem process projecting from the cell body. From this primary shaft, the single axon and various long, branched, tapering dendrites emerge (Brehmer 2021).

  12. 12.

    Important current data are available from the following sources: (1) cellular-resolution atlases of the human brain (e.g., Ding et al. 2016; Mai et al. 2016); (2) human brain anatomy using modified Brodmann or gyral annotation and gene expression data in 3D images (Allen Brain Atlas, https://human.brain-map.org/static/brainexplorer; https://atlas.brainmap.org/atlas?atlas=265297126#atlas=265297126&plate=112360888&structure=10390&x=40320&y=46976&zoom=-7&resolution=124.49&z=3; and, https://atlas.brainmap.org/atlas?atlas=138322605#atlas=138322605&plate=112360888&structure=10390&x=40320&y=46976&zoom=-7&resolution=124.49&z=3); (3) atlas of transcriptional features of the mid-gestational human brain (Miller et al. 2014); (4) segmentation of volumetric brain MRIs of infants (first 2 years of life; de Macedo Rodrigues et al. 2015); (5) connectomic study of adult human cerebral cortex on a petascale fragment (e.g., Shapson-Coe et al. 2021; available to peruse online and accessible with the Neuroglancer browser interface); (6) multimodal cell census and atlas of the mammalian primary motor cortex, including human data, and integrating neuronal multi-layered molecular genetic and spatial information with multi-faceted phenotypic properties (BRAIN Initiative Cell Census Network (BICCN) 2021); (7) the “BigBrain” model based on human cell body-stained and 3D-reconstructed sections as an anatomical brain model at a spatial resolution of 20 μm to be associated with ultra-highfield fMRI (with “possibility to measure laminar brain activity as well as identifying functional subdivisions of subcortical and cortical structures”; Kiwitz et al. (2022), and references therein; see also https://interactive-viewer.apps.hbp.eu/, https://ebrains.eu/, and https://ebrains.eu/service/voluba/); and (8) the methodological approach for studying synaptic and non-synaptic profiles of nearby axons and dendritic spines in Kasthuri et al. (2015), for example.

  13. 13.

    Immunohistochemical characterization of hypothalamic neurons and description of species differences in neuroanatomical subdivisions can be found in Saper (2012). Examples of Golgi-impregnated spiny neurons in hypothalamic nuclei obtained from other species were depicted by Cajal (1909–1911), and the history of neuroendocrinology since the descriptions in De humanis corporis fabrica by Vesalius can be found in Kreier and Swaab (2021). See data on pituitary alterations in humans in Baltazar-Gaytan et al. (2019).

  14. 14.

    For example, horizontal cells in layers I and II, double bouquet cells in layers II and III, chandelier and neurogliaform cells in layers II to IV, basket cells in layers III and IV, spiny stellate cells in layer II and aspiny stellate cells in layer IV, Martinotti cells in layer VI, fusiform stellate cells in layer VI, and a varied group of pleomorphic neurons identified as “modified pyramidal cells” in layer VI, among other varieties of non-spiny or sparsely spinous non-pyramidal heterogeneous cells found in the cerebral cortex of various species (see further data in Lodato and Arlotta (2015) and Kubota et al. (2016) for the complexity of these local cells; and, in addition, Braak 1980; Feldman 1984; Wahle 1993; Pearson 1995; Rudy et al. 2011; Leopold et al. 2019).

  15. 15.

    See also data on human cortical stimulus-related hemodynamic changes and the modulation made by task difficulty, arousal, and behavioral performance in Oelschlägel et al. (2022) and Burlingham et al. (2022), respectively. Commentaries on results from alert-behaving monkeys can be found in Sirotin and Das (2009) further discussed in Leopold (2009).

  16. 16.

    “In humans, the motor nucleus of the facial nerve is the largest of all motor nuclei of the brainstem... The comparative anatomy of the facial musculature and of the central nervous apparatus that controls facial movements suggests that, in some primates, group size, facial motor control, and primary visual cortex evolved with the same pattern... Species living in relatively large social groups tend to have relatively large facial motor nuclei, and species with enlarged facial nuclei and facial mobility have rather large primary visual cortices... Great apes and humans have facial motor cortices that are thicker and richer in local circuitry, their facial movements have the highest degree of dependence on the primary motor cortex... (and) more pronounced direct cortico-facial projections” (Müri 2016 and references therein).

  17. 17.

    As described by Holstege and Subramanian (2016), human speech needs the activation of two motor systems: one generates vocalization by activating the prefrontal - PAG - nucleus retroambiguus (NRA)–motoneuronal pathway, and the other modulates vocalization into words and sentences by activating the motor cortex and the corticobulbar fibers to specific muscles. The PAG has a central role in the “emotional motor system” when vocalizations also express emotions (e.g., crying and laughter in humans). “The PAG receives strong projections from higher limbic regions and from the anterior cingulate, insula, and orbitofrontal cortical areas. In turn, the PAG has strong access to the caudal medullary NRA. The NRA is the only cell group that has direct access to the motoneurons involved in vocalization, i.e., the motoneuronal cell groups innervating soft palate, pharynx, and larynx as well as the diaphragm, intercostal, abdominal, and pelvic floor muscles. Together they determine the intraabdominal, intrathoracic, and subglottic pressure, the control of which is necessary for generating vocalization. Only humans can speak, because, via the lateral component of the volitional or somatic motor system, they are able to modulate vocalization into words and sentences” (Holstege and Subramanian 2016).

  18. 18.

    The morphology, neurochemical profile, and number of affected SN pars compacta neurons relate the pathophysiology of Parkinson’s disease (PD; Gibb and Lees 1991). Recently, single-nucleus RNA-sequencing profiling of human SN pars compacta dopaminergic neurons identified a population that is selectively vulnerable and degenerates in PD (Kamath et al. 2022). These neurons have a unique, very large, dense, and widely spread axonal arbor architecture targeting the neostriatum (Matsuda et al. 2009), which might impose a chronic high energetic demand and likely damaging oxidative stress (Bolam and Pissadaki 2012; see additional comments in Giguère et al. (2018) and recent findings in an animal model in Ferreira et al. (2020)). There are various intracellular molecular changes in human SN dopaminergic neurons critical to the neurodegeneration and symptomatic manifestations of PD (Halliday et al. 2005), but the loss of neurons in this disease is not confined to the SN. Morphological and functional impairments include motor and nonmotor circuits, and PD progression alters the olfactory bulb, other brainstem areas, temporal mesocortex, and neocortical areas (Braak et al. 2003; Bolam and Pissadaki 2012; Giguère et al. 2018). For example, the pedunculopontine nucleus in the ponto-mesencephalic tegmentum is part of the mesencephalic locomotor region and may be involved in sleep and cognitive disturbances (French and Muthusamy 2018) at different stages of PD progression (Braak et al. 2003; see also Rietdijk et al. 2017). Many patients also suffer from hallucinations and delusions.

  19. 19.

    The functional participation in cortical circuits is not restricted to a few amygdaloid nuclei. Due to multimodal input convergence from multiple parallel circuits, directly or indirectly, amygdaloid nuclei of different origins and in “a degree of mosaic evolution” and function can connect higher-order cortical areas (Hirata et al. 2009; Yilmazer-Hanke 2012; Kedo et al. 2018). For example, “being part of an organized neural network that projects to the bed nucleus of the stria terminalis and to various hypothalamic and brainstem nuclei, the CeA and MeA subnuclei also participate in social and defensive reactions against innate and learned threats with neuroendocrine, behavioral, and sympathetic/parasympathetic responses to fearful/defensive and stressful stimuli... The MeA projects to the periallocortical, paleocortex, and archicortex, as well as to the insular agranular cortex and ventromedial prefrontal cortex... These data indicate that the MeA also participates, although with varied magnitude, in parallel circuits with different parts of the evolved neocortex for social and emotional processing in our species” (Rasia-Filho et al. 2021 and references therein)

  20. 20.

    For the description of putative pyramidal neurons in the human basal nucleus of Meynert; see Mesulam et al. (1983), Saper and Chelimsky (1984), Morrison et al. (1987), and Liu et al. (2015). In rats, see the description of pyramidal-like neurons and Fig. 1F in Nakajima et al. (1985).

  21. 21.

    The CA3 pyramidal neurons are critical for encoding, storage, and recall of associative memory along the mouse transverse (or proximodistal) CA3 axis (Sun et al. 2017), showing distinct projections to CA1 cells for processing nonspatial information in rats (Nakamura et al. 2013). Mice CA3 pyramidal neurons show heterogeneous morphology, afferent connectivity, and electrophysiology. Regular-spiking pyramidal neurons have complex spines (thorny excrescences) that receive monosynaptic mossy-fiber input and were found at any position along the radial axis of the CA3 stratum pyramidale. On the other hand, intrinsically bursting pyramidal neurons lack thorny excrescences (“athorny” cells), receive no or very few mossy-fibers input, and preferentially occupy distal CA3 parts closer to stratum oriens (Hunt et al. 2018).

  22. 22.

    For relevant data, see Baloyannis et al. (2001), Fiala et al. (2002), Penzes et al. (2011), Jiang et al. (2013), Overk and Masliah (2014), Maiti et al. (2015), Herms and Dorostkar (2016), Hrvoj-Mihic et al. 2017, Forrest et al. (2018), Chidambaram et al. (2019), Runge et al. (2020), Bączyńska et al. (2021), and Parajuli and Koike (2021), for example.

References

Download references

Author information

Authors and Affiliations

Authors

Editor information

Editors and Affiliations

Rights and permissions

Reprints and permissions

Copyright information

© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Check for updates. Verify currency and authenticity via CrossMark

Cite this chapter

Renner, J., Rasia-Filho, A.A. (2023). Morphological Features of Human Dendritic Spines. In: Rasia-Filho, A.A., Calcagnotto, M.E., von Bohlen und Halbach, O. (eds) Dendritic Spines. Advances in Neurobiology, vol 34. Springer, Cham. https://doi.org/10.1007/978-3-031-36159-3_9

Download citation

Publish with us

Policies and ethics