| # | Study | Design | Species | Sample size | Key findings (abridged) | OCEBM | RoB | Access |
|---|---|---|---|---|---|---|---|---|
| 1 | Dikeçligil & Gottfried, 2024 | narrative review | human | NA (narrative review of 10–15 years of human olfaction research) | Humans have nearly 900 OR genes, only ~400 still functional; combinatorial code (each odorant binds several ORs, each OR binds several odorants); Mainland et al. (2014): OR10G4 genotype explains 15% of the variance in intensity and >10% of the variance in valence of guaiacol; 63% of orphan ORs deorphanised… | 5 | NA | TV |
| 2 | Branigan & Tadi, 2023 | narrative review (book chapter) | human | NA (chapter synthesised from 32 references) | Each OSN expresses one receptor type, and one odorant binds several receptor types; an estimated 1,100–1,200 glomeruli per olfactory bulb in humans; the fetus discriminates odours in amniotic fluid from about week 30; the olfactory bulb matures until about age 2 | 5 | NA | TV |
| 3 | Barnum & Hong, 2022 | narrative review (Primer, cross-species comparison) | multi | NA (review; mainly mouse and Drosophila, extending to zebrafish, locust, amphibians, C. elegans) | Number of functional odorant receptor genes: Drosophila ~60; C. elegans ~1,200; zebrafish ~160; mouse ~1,100; human ~400 — gene number does not directly predict olfactory ability; OSN→glomerulus convergence ratio ~25:1 in flies, ~250:1 in mice; GABAergic lateral inhibition performs normalisation… | 5 | NA | TV |
| 4 | Torske et al., 2022 | coordinate-based ALE meta-analysis (fMRI/PET; PROSPERO CRD42020157138, PRISMA) | human | 81 studies, 108 contrasts, 1,053 foci, 1,744 participants; pleasant 38 contrasts/302 foci (623 participants);… | All odors: largest cluster left amygdala/piriform (k = 1,163; −22, 0, −20; ALE 0.127) and right amygdala/piriform (k = 886); also OFC, bilateral insula, parahippocampal, left caudate; Pleasant: bilateral amygdala/piriform (right k = 247; 26, 2, −18; left k = 213), OFC… | 3 | some concerns | TV |
| 5 | Zhou et al., 2021 | resting-state fMRI + iEEG (cross-sectional) | human | fMRI: 25 healthy participants (mean age 25.5 ± 1.2; 14 female); iEEG: 8 epilepsy patients (3 female), 72 electrode pairs | Main effect of region on hippocampal connectivity: left F(3,72) = 25.80, P = 1.92e-11; right F(3,72) = 16.37, P = 3.29e-8; olfactory > visual, > auditory, > somatosensory (FDR P < 0.05); all olfactory subregions (AON, olfactory tubercle, frontal piriform, temporal piriform) were stronger… | 5 | some concerns | TV |
| 6 | Gu et al., 2025 | systematic review (self-described; methods not accessible) | human | NR (115 references according to Crossref / 120 according to OpenAlex) | 391 functional human odorant receptors classified; 402 OR–ligand pairs compiled | 5 | NA | Abs |
| 7 | Wu et al., 2024 | narrative review | NA | NA | ORs are the largest family of chemosensory receptors, belonging to class A GPCRs; new structures of OR51E2 and consensus OR52 reveal features distinguishing ORs from other class A GPCRs (abstract does not specify the features) | 5 | NA | Abs |
| 8 | Kurian et al., 2021 | narrative review | multi | NA (review; cites Nara 2011: ~3,000 OSNs × 125 odours ≈ 375,000 pairs; Xu 2020: ~10,000 cells; Pfister 2020:… | Number of intact OR genes ~400 in humans, ~1,100 in mice, ~2,000 in elephants; each mature OSN expresses one allele of one OR gene; OR expression levels differ by up to 300-fold; up to 98.9% of mouse ORs are expressed; Bushdid et al. (2014): humans discriminate at least 1.72 trillion odours… | 5 | NA | TV |
| 9 | Reactome R-HSA-381753 (Caudy & May; 2008/2021) | curated pathway database | human | NA (many reactions inferred from mouse/rat orthologues) | The human genome has ~857 potential OR genes, ~394 possibly functional; observed expressed numbers: 143 (Durante 2020, scRNA-seq), 273 (Verbeurgt 2014), 346 (Olender 2016); Cascade: odour → OR → GNAL (Gαolf) GDP/GTP exchange → ADCY3 produces cAMP → CNG channel (2 CNGA2 : 1 CNGA4 : 1… | 5 | NA | TV |
| 10 | DeMaria & Ngai, 2010 | narrative review | multi | NA (review; data mainly from mice, plus human genetics) | Humans have about 700 OR genes, about half still functional; rodents >1,200 genes, about 2/3 functional; fish 40–140 intact genes; pseudogenes ~25–30% in rodents, ~50% in humans; Mouse: ~5–10 million OSNs, ~1,800 glomeruli per olfactory bulb → convergence on the order of 10³ axons… | 5 | NA | TV |
| 11 | Kehl et al., 2024 | single-unit (microwire) recording in epilepsy patients | human | 17 epilepsy patients (12 female; age 22–60, mean 41.3 ± 11.7), 27 sessions; 2,416 units (1,292 single units):… | Proportion of odour-modulated neurons: PC 39.5 ± 4.7% (Z = 3.6, P = 0.00029); amygdala 19.5 ± 2.7% (P = 1.9 × 10⁻⁵); EC 14.2 ± 3.2% (P = 0.049); hippocampus 12.1 ± 1.9% (P = 0.0019); PHC 5.31 ± 1.4% (P = 0.78); Decoding of odour identity (200 neurons, chance 6.25%) had… | 5 | some concerns | TV* |
| 12 | Sagar et al., 2023 | fMRI dense sampling (RSA + encoding model) | human | 3 healthy participants (2 female aged 23–24, 1 male aged 24; 1 participant excluded); ≥ 4,320 trials/participant, 12,720 odour pairs/participant, ~18… | Perceptual RSA: OFC rp = 0.120, AMY 0.059, PirT 0.035 (P < 0.0001), PirF 0.005 (ns); perceptual > molecular in OFC (rm 0.077; P < 0.0001) and PirT (rm 0.024; P = 0.031); Fine code > coarse code: OFC rf 0.083 vs rc 0.063 (P < 0.0001), AMY 0.032 vs 0.024 (P =… | 5 | some concerns | TV |
| 13 | Sagar et al., 2026 | fMRI dense sampling (reanalysis of Sagar 2023 data; RSA + SVR decoding +… | human | 3 healthy participants (2 female aged 23–24, 1 male aged 24); S1 4,560 trials, S2/S3 4,320 trials; decoding N = 4,400 trials;… | Valence RSA significant in all 4 regions: OFC β = 0.569 (P < 0.001), vmPFC 0.495 (P = 0.001), AMY 0.356 (P = 0.031), PirC 0.331 (P = 0.027); salience in AMY, OFC, vmPFC but not in PirC (β 0.169, P = 0.182); Trial-wise decoding (N = 4,400, P < 0.001): OFC… | 5 | some concerns | TV |
| 14 | Zhang et al., 2024 | narrative review (mini review) | mouse | NA (review of mouse studies) | The OB projects directly to 9 olfactory cortical areas (AON, OT, aPir, pPir, LEC, nLOT, ACo, PLCo, TT); mitral cells project to all areas, tufted cells mainly to anterior areas; after go/no-go training, the number of activated glomeruli increases and persists for up to 5 weeks (Abraham 2014); mitral synchrony encodes… | 5 | NA | TV |
| 15 | Lyons & Gottfried, 2025 | narrative review / theoretical opinion | multi | NA (theoretical review; human, rodent, fly and computational-model data) | Proposed microcircuit: piriform sends predictions down to granule cells via feedback; granule cells suppress the predicted part of the input; mitral cells carry prediction errors up to piriform (based on Koulakov & Rinberg 2011; Grabska-Barwińska 2017); Piriform resembles the hippocampus (three-layered paleocortex… | 5 | NA | TV* |
| 16 | Fulton et al., 2024 | narrative review (fly–mouse comparison) | multi | NA (review of Drosophila and mouse, including fly larvae and rats) | Number of receptors: 23 ORs in fly larvae, 62 in adult flies, ~1,300–1,500 in mice; ~54 glomeruli in the fly AL, ~1,800 in the mouse OB; ~2,200 Kenyon cells; each KC receives 1–7 uPNs, each uPN projects to 50–300 KCs; connectomics shows uPN–KC connectivity is non-random (uPNs for food odours… | 5 | NA | TV |
| 17 | Su et al., 2009 | narrative review | multi | NA (review of mammals and insects, mainly Drosophila) | Mammals have 250–1,200 functional OR genes depending on species; 15–78% of ORs are pseudogenes; all human V2Rs have become pseudogenes; mice ~15 TAARs, ~200 V1Rs, ~100 V2Rs; insects 60–340 Or genes, ~60 IRs; Hallem & Carlson (2006): ~75% of Drosophila antennal Ors × 110 odours — tuning is a continuum from… | 5 | NA | TV |
| 18 | Kontaris et al., 2020 | narrative review | multi | NR (narrative review; selection criteria and number of sources not stated) | Villemure 2003: 5 minutes of pyridine induced negative mood and mild anxiety; commercial perfume induced positive, calm mood; the olfactory pathway bypasses the thalamus: OB → AON, OT, coA, aPCX/pPCX, LEC; coA → BLA; BLA ↔ pPCX reciprocally connected; optogenetic activation of BLA modulates… | 5 | high | TV |
| 19 | Soudry et al., 2011 | narrative review | human | NR (review; 44 references 1969–2010); data cited: Deems 1991 374 patients with olfactory disorders vs… | Table 1: amygdala, ACC, insula, OFC are involved in olfaction, emotion, depression and schizophrenia alike — the “shared substrate” argument; Deems 1991: 374 patients with olfactory disorders vs 362 age- and sex-matched controls had significantly higher Beck Depression Inventory scores… | 5 | high | TV |
| 20 | Saive et al., 2014 | narrative review (structured by experimental paradigm, focused on neuroimaging) | human | NR (review); number of neuroimaging studies: 4 standard recognition, 2 multimodal, 2 autobiographical | Engen & Ross 1973: odour recognition lower than picture recognition at short intervals (73% correct) but higher than pictures after 4 months; however Murphy 1991, Larsson 1997, Olsson 2009 show that odours are forgotten; Royet 2011 (fMRI): correct recognition activates hippocampus/parahippocampal, errors… | 5 | some concerns | TV |
| 21 | Kadohisa, 2013 | narrative review (mini review) | multi | NR | aPC performs pattern separation (encodes mixture identity separately from components), pPC encodes similarity; aPC neurons adapt quickly but still detect a new odour against a background of an old odour (Kadohisa & Wilson 2006a); the MOB projects to the anterior cortical nucleus, medial nucleus and periamygdaloid… | 5 | high | TV |
| 22 | Yang et al., 2025 | diffusion MRI tractography (cross-sectional) | human | 27 participants scanned (14 male, 13 female; mean age 25.76 ± 4.01), 2 males excluded → n = 25 analysed; 8 with hyposmia | ANOVA of endpoints/voxel across 8 nuclei: F = 9.3057, df = 7, p = 6.6 × 10⁻¹⁰; ranking MeA > CeA > ACo > PAC > LA > BLA > BMA > PCo; K-means: PCo, BLA, LA, BMA grouped in the same cluster in 100% of iterations (permutation p<0.05) | 5 | some concerns | TV |
| 23 | Noto et al., 2021 | resting-state fMRI + connectivity-based parcellation (k-means) + literature review | human | 25 healthy adults (14 female), mean age 25.5 (SE 1.2); data from Zhou et al. 2019 | k=2: olfactory nuclei grouped together spontaneously — right z = 4.27, P = 2.0 × 10⁻⁶; left z = 4.79, P = 1.7 × 10⁻⁶; k=3: three clusters matching the three anatomical nuclei MeA/CoA/PAC, minimum z = 6.05 | 5 | some concerns | TV |
| 24 | Buchanan et al., 2003 | lesion case-control (neuropsychological) | human | 20 patients with unilateral temporal lobectomy (11 left LTL, 9 right RTL) + 1 patient SM046 with amygdala lesion… | Odour–name matching: false alarms differed between groups F(2,37) = 5.6, p = 0.007, η² = 0.23; LTL vs NC p = 0.006; LTL vs RTL p = 0.03; d′ NC 0.48, LTL 0.15, RTL 0.27, SM046 −0.11; Pooled lobectomy vs NC: false alarms t(38) = 2.3, p<0.05; d′ t(38) = 1.7, p<0.05 (name matching);… | 4 | some concerns | TV |
| 25 | Mouly & Sullivan, 2010 | narrative review (book chapter) | multi | NR (review book chapter; mainly rat data) | Odour fear in adult rats extinguishes more slowly than tone/visual fear (Richardson 2002); BLA lesions before/after training abolish acquisition and expression (Cousens & Otto 1998); Sevelinges 2004: after odour fear conditioning, field potential amplitude in the cortical amygdala increased persistently for 24… | 5 | some concerns | TV |
| 26 | Hakim et al., 2019 | narrative review | multi | NR | Morrison 2015 (M71 mice): odour fear increased the number of neurons expressing the M71 receptor in the epithelium and the cross-sectional area of M71 glomeruli; odour-specific fear extinction reversed the changes; Kass 2013: sensory neurons expressing the specific receptor increased ~twofold after 3 weeks, glomerular output increased | 5 | high | TV |
| 27 | Li & Wilson, 2023 | narrative review (theoretical framework) | multi | NR (except Arnold 2020: connectome of ~800 people) | You 2022 (human, fMRI): >20% of piriform voxels originally tuned to non-CS odours shifted to tuning to a similar CS+ on day 9, not on day 1 — piriform plasticity increases over time; Chen 2011 (rat): generalised fear → expanded piriform receptive fields + generalised freezing… | 5 | some concerns | TV |
| 28 | Sullivan et al., 2015 | editorial (introducing a 25-article Research Topic) | multi | NR (editorial summarising 25 member articles; topic page lists 26 articles) | Pool 2014: after appetitive conditioning, people did not consciously discriminate two similar odours but responded faster, sniffed more deeply and showed higher skin conductance to the reward-paired odour → learning below the threshold of awareness; Ferdenzi 2014: repeated odour exposure causes affective habituation in pleasantness… | 5 | NA | TV |
| 29 | Tong et al., 2014 | narrative review (mechanisms; rodent data) | multi | NR (review); Figure 5 data: t(8) suggests 9 animals | More reward pairings (3, 6, 12) sharpen the generalisation gradient; variable rewarded odours make the gradient track the environmental distribution (statistical learning); aversive learning broadens the gradient (Chen 2011); After 12 massed pairings, the gradient gradually flattens with delays of 2–1440 minutes:… | 5 | some concerns | TV |
| 30 | Patin & Pause, 2015 | review (synthesis of fMRI/PET studies in healthy people; abstract describes it as “comprehensive”, criteria… | human | NR (number of studies not accessible) | Pleasant odours → bilateral or left amygdala activation; Unpleasant odours → right amygdala activation (authors’ interpretation: survival role of the right amygdala) | 5 | NA | Abs |
| 31 | Herz et al., 2004 | fMRI (within-subject, 4 conditions) + behavioural ratings | human | 5 healthy right-handed women | Recall evoked by the PS odour activated the amygdala and hippocampal region significantly more than the PS picture, the control odour and the control picture; Saive 2014 specifies: parahippocampal gyrus, amygdala, middle occipital gyrus | 4 | high | Abs |
| 32 | Hackländer et al., 2019 | theoretical review + mini-meta-analysis | human | 13 studies, 14 experiments (samples from Germany, the Netherlands, Sweden, the UK, the USA; mainly students); total N not stated | Memory age: decade × modality interaction F(2,8) = 16.20, p = .002, ηp² = .80; first decade: odour M = .51 vs other .22, t(4) = 5.63, p = .005; Emotional intensity: M d = .47, t(9) = 2.33, p = .022; feeling of being brought back: M d = .47, t(6) = 2.36, p = .028 | 3 | some concerns | TV* |
| 33 | Herz, 2026 | perspective (theoretical viewpoint) | human | NA (no new data) | Herz 2003: heart rate and skin conductance decreased when participants believed the odour was “relaxing” and increased when they believed it was “energising”, regardless of whether a real odour or no odour was present; the mood effects of aromatherapy are not pharmacological but due to learned associations (Herz 2009); people who do not discriminate natural odours… | 5 | high | TV |
| 34 | Dejou et al., 2026 | survey (human) + animal experiment (behaviour, USV, CldU/cFos, optogenetics, cFos map of 27… | multi | Human survey n = 647 (174 male, 473 female; aged 18–85, M = 45.4, SD = 17.2); 272 male and female C57BL/6J mice (e.g.… | Humans (n = 647): 73.1% of earliest odour memories came from events repeated > 5 times; mean pleasantness 7.1/9; joy and pleasantness highest; 82% still re-encounter the odour periodically; those whose descriptions were reported were more pleasant than those not reported, t(98.9) = 2.01, p = 0.024; PLAY-O emitted more USVs than CTRL-O (count p =… | 5 | some concerns | TV |
| 35 | Yang et al., 2021 | iEEG (intracranial LFP) + SVM decoding, olfactory Sternberg task | human | 8 patients with drug-resistant epilepsy (3 male, 5 female; aged 27–57); PCx electrodes in 7, AH in 6, both in 5; 50–60… | Theta–gamma modulation strength higher on correct-identity trials (PCx t6 = 4.66, p = 0.004; AH t5 = 5.74, p = 0.002); Theta phase encodes sequence position: phase separation correct vs incorrect PCx 2.2 ± 0.3 vs 0.3 ± 0.3 (t5 = 5.97, p = 0.002); AH 2.8 ± 0.2 vs 0.3 ± 0.3 (t5 = 7.32, p… | 4 | some concerns | TV |
| 36 | Szychowska et al., 2025 | systematic review + meta-analysis (PRISMA 2020, not preregistered) | human | 24 articles in the review; meta-analysis k = 7 (olfactory vs visual); mean 68 participants/study (SD 104, median… | Meta-analysis olfactory vs visual (k = 7): µ = −0.48, CI95 [−0.77, −0.19], z = −3.22, p = 0.0013 → olfactory spatial memory is poorer than visual; individual-study SMDs −1.13 to −0.01, all negative; Heterogeneity Q(6) = 23.03, p = 0.0008; τ² = 0.095; I² = 71.09%; 95% prediction interval… | 3 | some concerns | TV |
| 37 | Huang et al., 2026 | systematic review (PRISMA 2020; PROSPERO CRD42024559585; narrative synthesis + frequency counting… | human | 164 studies (4,313 records → 2,374 screened → 339 full texts); OB/OS: 63 studies, 3,832 patients; volume… | 45/50 studies (90%) with a control group showed reduced OBV and/or OSD; exception Covid-19 with increased OBV by Box-frame (Sherif 2022: 83.6 vs 30.5 mm³; Abdou 2023: 81.3 vs 31.8 mm³), attributed to oedema/inflammation and methodology; OBV–function: 20/23 (86.9%) significant positive correlation… | 3 | some concerns | TV* |
| 38 | Smitka et al., 2012 | cross-sectional (manually traced volumetric MRI + Sniffin' Sticks) | human | 117 healthy volunteers, aged 19–77 (mean 37), 62 female / 55 male | Reference volumes: right HC 3.29 cm³ (SD 0.47); left HC 3.15 cm³ (SD 0.47); right AG 1.60 cm³ (SD 0.31); left AG 1.59 cm³ (SD 0.30); HC and AG volumes decrease with age, markedly more on the right | 3 | some concerns | Abs |
| 39 | Baudry, 2020 | narrative review / essay | NA | NA (51 references) | Examines the popular claim that Proust’s involuntary odour/taste-evoked memory anticipated episodic memory (Tulving); focuses on the role of the hippocampus in encoding time and the components of an episode, especially olfactory information | 5 | NA | Abs |
| 40 | Buck & Axel, 1991 | molecular cloning (degenerate PCR + cDNA library screening + Southern/Northern blot +… | rat | Sprague-Dawley rats (number of animals NR); 64 PCR bands; 36 partially sequenced clones = 18 distinct sequences;… | Isolated 18 distinct members of a multigene family encoding 7TM proteins (each hydrophobic segment 19–26 aa), conserved motifs GN (TM1), W (TM4), Y (TM5), NP (TM7), DRY after TM3; Expression only in olfactory epithelium: Northern 2 bands ~2 kb and ~5 kb; no signal in brain, retina… | 5 | NA | TV |
| 41 | Keller et al., 2007 | heterologous expression screen (Hana3A + CRE-luciferase) + genotyping + psychophysics… | human | 412 raters (218 female, 194 male; median age 34, 19–75); 255 RT/RT, 100 RT/WM, 10 WM/WM, RT/P79L 10… | Of 337 functional human ORs, only OR7D4 responded clearly to androstenone 30 µM; OR7D4 RT responded only to androstenone and androstadienone among 66 odours + 2 solvents; WM responded to none; SNPs R88W and T133M share frequency 0.154, in complete linkage (r² = 1, D′ = 1) forming the allele… | 3 | some concerns | TV* |
| 42 | Billesbølle et al., 2023 | cryo-EM + mutagenesis + cell assay (GloSensor cAMP) + molecular dynamics + phylogenetic… | human | 16,113 movies; 8,884,130 initial particles; 204,438 particles for local refinement; MD 5 × 1 µs per condition;… | First cryo-EM structure of a human OR (OR51E2–miniGs–Gβ1γ2–Nb35) in the active state, global map 3.1 Å; Propionate binds in a small pocket (31 ų), fully occluded, between TM3–TM6, with no contact with TM2/TM7; the carboxylate is anchored by an ionic interaction with R262^6x59… | 5 | NA | TV* |
| 43 | Stephan et al., 2012 | animal experiment (global knockout + conditional OMP-Cre) + EOG + suction pipette… | mouse | EOG dose-response/latency 9 WT / 10 het / 7 KO; termination n = 16/19/17 (10⁻⁴, 10⁻³ M) and 9/10/9 (concentration… | Nckx4⁻/⁻ mice have normal olfactory epithelium (OMP/GAP-43, EdU, Caspase-3, ACIII, CNGB1b, glomerulus number) and EOG peak amplitude, latency and time-to-peak like WT at all concentrations — but the termination τ is 2–3 times longer; heterozygotes are also slower (gene-dose effect); Single cells (IBMX 0.3… | 5 | some concerns | TV |
| 44 | Reisert & Zhao, 2011 | narrative review (Perspective) | multi | NA (synthesis; Figure 3E from Song 2008: n = 19–49 OSNs) | Cl⁻ current through Ano2 carries up to 90% of the odour-induced receptor current; rodent breathing frequency ~2–3 Hz at rest, up to ~10 Hz during exploratory sniffing; Action potentials fire only during the rising phase of the response; latency can be ~30 ms, maximum frequency ~200 Hz; narrow dynamic range ~2 log units; at high concentrations the number of spikes… | 5 | some concerns | TV |
| 45 | Nagayama et al., 2014 | narrative review (anatomy–function of olfactory bulb circuits) | multi | NA (review) | Marker proportions in PG cells (mouse): CR 30.0–44.0%, TH 12.6–20.0%, CB 9.8–15.0%; CR, CB, TH are mutually exclusive (TH⁺ are type-I, CB⁺/CR⁺ are type-II); GAD65/67 in > half of juxtaglomerular cells; TH⁺/GAD67⁺ sSA cells have axons ~1 mm, dendrites contacting up to ~50 glomeruli… | 5 | NA | TV |
| 46 | Gottfried, 2010 | narrative review (theoretical review) | multi | NA (review) | Trained wine tasters reliably identify only ~3 components in a mixture; odour perception is configural; cooked clam broth contains 49 odorants; Habituation to A+B in rats: the olfactory bulb decreases with A or A+B, APC decreases only with A+B →… | 5 | NA | TV* |
| 47 | Wilson & Sullivan, 2011 | narrative review (animal/electrophysiological perspective) | multi | NA (review); LEC 45/128 units odour-responsive; optogenetics ~500 cells (≈0.5% of piriform) | Piriform neurons require co-activation of several glomeruli to fire (one glomerulus is not enough); neighbouring neurons respond to different odours as often as to the same odour, with no spatial structure at any scale; Optogenetic activation of ~500 cells (≈0.5% of the population… | 5 | NA | TV* |
| 48 | Gottfried et al., 2006 | fMRI (olfactory cross-adaptation) | human | NR (not accessible) | Double dissociation in human piriform: PPC encodes odour quality but not structure; APC encodes structure but not quality (abstract); the structural code in APC is interpreted as reflecting sensory fidelity from the olfactory bulb; the quality code in PPC independent… | 4 | some concerns | Abs |
| 49 | Howard et al., 2009 | high-resolution fMRI (3T, 1.72 × 1.72 × 2 mm) + MVPA (split-half correlation, MDS +… | human | Total 20 participants, aged 22–35: Exp. 1 n = 6 (4 female); Exp. 2 n = 4 (3 female); descriptor questionnaire n = 10 (8 female); 24 runs/participant… | Univariate: mean PPC activity did not differ between odours in 5/6 participants; classification by mean activity was not above chance in any region (Exp. 2 P = 0.16–0.44); sniff parameters could not classify (Exp. 1 all P > 0.18); Exp. 1: MVPA decoded identity… | 4 | some concerns | TV* |
| 50 | Zelano et al., 2011 | high-resolution fMRI + MVPA (odour search task, pattern correlation) | human | 13 recruited (6 female; aged 19–23), 1 excluded → n = 12; 6 blocks × 32 trials + localiser block of 18 trials | Behaviour: A+B trials less accurate than A (T11 = 4.39, P = 0.001) and B (T11 = 3.96, P < 0.002); congruent trials had fewer errors (T11 = 3.35, P < 0.006) and were faster (T11 = 3.01, P < 0.01) than incongruent trials; APC: pre-odour patterns same-target >… | 4 | some concerns | TV |
| 51 | Jiang et al., 2017 | iEEG (depth electrodes in epilepsy patients) + time–frequency analysis + SVM decoding +… | human | 9 patients recruited, 2 excluded → n = 7 (6 female; aged 19–61); PPC–hippocampus PLV pooled across 4 participants; intensity ratings… | Odours increased theta power locked to sniff onset in each patient (peak Z 4.2–10.4), peak at 300–800 ms (mean 447 ms), mean peak frequency 4.53 Hz; group theta odour > air Wilcoxon Z = −2.37, p = 0.018; Theta strongest medially and anteriorly: regression of electrode position… | 4 | some concerns | TV |
| 52 | Anderson et al., 2003 | fMRI (event-related), 2×2 intensity × valence design | human | 16 healthy participants (8 female, 8 male) — according to the secondary source AuntMinnie, not checked against the original | Amygdala activation was linked to odour intensity, not valence (abstract); OFC regions were linked to valence independently of intensity (abstract) | 4 | some concerns | Abs |
| 53 | Grabenhorst et al., 2007 | 3T event-related fMRI + trial-by-trial psychophysical ratings (parametric modulation) | human | 14 healthy participants (7 male, 7 female, mean age 26), selected on finding indole unpleasant and the mixture pleasant; 4… | Pleasantness F(2,26) = 41.8, p ≪ 0.001: jasmine ≈ jasind (p > 0.3) > indole; intensity did not differ (F(2,26) = 2.43, p > 0.1); medial OFC [4, 46, −14] (Z = 2.41) correlated positively with pleasantness (r = 0.88): jasind > jasmine (p = 0.031) > indole (p = 0.014) —… | 4 | some concerns | TV |
| 54 | Bao et al., 2019 | 3T fMRI + behaviour (mental navigation task in a 2D odour space; quadrature… | human | 52 consented (40 female) → 34 reached ≥70% discrimination → 25 scanned: n = 25 (22 female; aged 18–37, mean 24.3) | Prediction accuracy above the 50% chance level on every day (no difference between days, F(3.41, 81.91) = 1.41, p = 0.24); orientation index improved from day 1 (t24 = −2.17, p = 0.04); mean odour discrimination accuracy 77%; vmPFC (5 mm sphere centred at 6, 44, −10) showed modulation… | 4 | some concerns | TV |
| 55 | Bushdid et al., 2014 (with Meister 2015; Gerkin & Castro 2015) | psychophysics (3-AFC odd-one-out, double-blind) + mathematical extrapolation (sphere packing); with two… | human | 28 recruited, 2 excluded → n = 26 (17 female; median age 30, 20–48); 264 tests/participant (260 mixture pairs = 13 types ×… | Discrimination decreased as overlap increased: ≥ half of participants discriminated pairs with overlap < 75%; no one discriminated > 90%; 227/260 pairs above chance, 148 pairs with ≥ 14/26 participants correct; Regression: y = −0.81x + 91.45 (by participant) and y = −0.77x + 94.22 (by… | 4 | high | TV* |
| 56 | Lee et al., 2023 | machine learning (graph neural network) + psychophysics with a trained sensory panel;… | human | Training ~5,000 molecules (138 labels, GoodScents + Leffingwell); validation 400 new molecules → 323 rated;… | Selected GNN had mean CV AUROC 0.89; panel test-retest R = 0.80 (n = 15); Prospective validation: the GNN was closer to the panel mean than the median rater for 53% of molecules (RF 41%); outperformed the median rater on 30/55 labels (55%); outperformed RF (paired t-test p = 3.3e-7);… | 3 | some concerns | TV* |
| 57 | Yeshurun & Sobel, 2010 | narrative review / theory | human | NA (review) | Multidimensional olfactory input (≈1,000 receptor types according to the notes) yields an almost one-dimensional percept; what people consistently report is whether an odour is pleasant or not (abstract); People can discriminate countless odours but can name very few (“an odour is not worth a thousand… | 5 | NA | Abs |
| 58 | Sela & Sobel, 2010 | narrative review / theory (citing the group’s data) | human | NA (review); data cited: Lorig 1992 — 3/93 people noticed the odour manipulation | Very high sensitivity: ethyl mercaptan detected at 0.2–0.009 ppb; isoamyl mercaptan 0.77 parts per trillion; discrimination of concentration differences of ~7%; yet only 3/93 people noticed a vanilla/lavender odour in the room (Lorig 1992); Ethyl mercaptan in propane at 0.5 ppm ≈ 57,000 times… | 5 | NA | TV |
| 59 | Keller, 2011 | narrative review / theory (philosophy–cognitive neuroscience) | human | NA (review) | Olfaction lacks covert spatial, object and feature attention (by the author’s definitions) but has temporal and cross-modal attention; odour experience is sampled about once every ~1.6 s owing to sniffing; fMRI evidence for attentional modulation in piriform is split (2/4 studies:… | 5 | NA | TV |
| 60 | Zald & Pardo, 1997 | PET H2(15)O rCBF, within-subject (odour vs odourless control) + post hoc ROI correlation | human | 12 women (aged 19–49); n = 10 for the sulfide subtraction (2 excluded for unpleasantness scores < 4); 8 additionally scanned… | Sulfide very unpleasant (1.3; SD 1.2) and very intense (8.6; SD 1.4); UPSIT less unpleasant (2.8; t(16) = 2.3; p < 0.04) and less intense (5.8; t(16) = 4.1; p < 0.001); Sulfide vs control: left amygdala (−34, −4, −11) Z = 5.2; right amygdala (26, −1, −14) Z =… | 4 | some concerns | TV |
| 61 | Gottfried et al., 2003 | fMRI + appetitive Pavlovian conditioning + reward devaluation by selective satiation (pre/post) | human | 13 healthy, hungry participants (according to Dolan 2007) | Olfactory associative learning (CS+ vs CS−) recruited the amygdala, rostromedial OFC, ventral midbrain, piriform, insula and hypothalamus; after selective satiation, responses to the CS+ predicting the devalued odour decreased in the left amygdala and in both rostral and caudal OFC; the CS+ predicting the non-devalued odour was unchanged | 4 | some concerns | Abs |
| 62 | Winston et al., 2005 | event-related fMRI, manually traced amygdala ROI, 2×3×3×2 ANOVA + additional psychophysical experiment | human | fMRI: 18 recruited, n = 17 analysed (12 female; aged 20–29); psychophysics: 17 recruited, n = 16 (13 female; aged 18–40) | Concentration × odour type interaction in the amygdala F(2.0, 31.7) = 4.7, p = 0.017; main effect of concentration F(1,16) = 11.1, p = 0.004; High − low difference: pleasant 0.14 % ± 0.061 (t(16) = 2.3, p = 0.017); unpleasant 0.23 % ± 0.058 (t(16) = 4.1, p < 0.001); neutral 0.0001 % ±… | 4 | some concerns | TV |
| 63 | Li et al., 2008 | multivariate fMRI (MVPA) + psychophysics (triangular test) + Pavlovian fear conditioning (electric shock),… | human | Not stated in the main text; secondary source (Ed Yong 2008): 12 volunteers | Before conditioning, discrimination was at chance; afterwards, tgCS+ vs chCS+ accuracy more than doubled (p ≤ 0.01, Wilcoxon), the CS− pair did not improve; intensity/valence/familiarity did not differ (p > 0.4); Posterior piriform: pattern correlation of the CS+ pair decreased after conditioning… | 4 | some concerns | TV* |
| 64 | Krusemark et al., 2013 | fMRI before/after anxiety induction (within-subject, no control group) + behaviour + PPI + DCM | human | 14 adults (aged 18–28; 8 male) for behaviour; fMRI n = 12 (2 excluded for head motion) | Anxiety rose from 15.64 (SD 19.06) to 40.36 (SD 28.25); t(13) = 4.99; p < 0.001; Valence of neutral odours decreased −0.59 → −1.82 (post < pre t(13) = −1.77, p < 0.05 one-tailed); change more negative than for negative odours t(13) = −3.03, p < 0.01; correlated with the increase… | 4 | high | TV |
| 65 | Iravani et al., 2021 | non-invasive olfactory bulb EEG/EBG + RSA (Exp 1); approach/avoidance behaviour on a force plate, registered… | human | Exp 1: 19 participants (28.88 ± 4.52 years; 7 female), 3 sessions, 540 trials/participant; Exp 2: pilot n = 21, main n = 44 (26… | Gamma (≈53–65 Hz) phase-coupled to beta from ≈250 ms after odour onset (t = 3.85, p < 0.006); gamma–valence RSA significant at ≈250–325 ms (r = 0.60, p < 0.010); beta–valence at ≈800 ms (r = 0.65, p < 0.014) | 4 | some concerns | TV |
| 66 | Root et al., 2014 | animal experiment — optogenetics (eNpHR3.0 inactivation; ArcCreERT2 + ChR2 activation), assay… | mouse | n = 3–11/group (e.g. CoA inactivation n = 11; tubercle n = 7; piriform n = 8; activation of TMT-labelled neurons n = 6, 2-PE… | Assay: TMT PI −83 ± 5.0; isopentylamine −81 ± 4; 4-methylthiazole −59 ± 8.1; peanut oil +50 ± 4.1; 2-phenylethanol +49 ± 4.8 (ANOVA p < 0.001); Inactivation of olfactory bulb terminals in CoA: TMT PI −65 ± 3.4 → −7.9 ± 8.4 (n = 11); 2-phenylethanol +48 ± 3.2 → −8.3 ± 2.1… | 5 | some concerns | TV |
| 67 | Chrea et al., 2009 | survey/psychometric: term selection (210 raters) + EFA (Study 1) + EFA/CFA (Study… | human | Term selection: 210 people; Study 1: 38 students (24 female), analysis of 19 (emotion) and 18 (quality) per set; Study 2: 317… | Study 1: 5 emotion factors (α 0.92–0.97) and 4 quality factors (refined, heavy, wholesome, sweet; α 0.82–0.96); between-set reliability r = 0.88; Canonical correlation emotion ↔ quality r = 0.83 (p < 0.001), 4 canonical variate pairs each ≈20 % of variance | 3 | some concerns | TV |
| 68 | Bensafi et al., 2002 | psychophysiology (skin conductance, heart rate) + subjective ratings in a separate session; factor analysis… | human | 12 healthy volunteers | The pleasantness factor correlated with heart-rate variation (He 2014: inverse — the more pleasant the odour, the more heart rate decreases/the less it rises); the arousal factor correlated with skin-conductance variation | 4 | high | Abs |
| 69 | Pause, 2012 | narrative review | human | NA (review; 1–2 imaging studies per area) | HLA-similar odours are processed faster and recruit more resources (Pause 2006, CSERP); the odour of a sister (not consciously discriminable) increased rCBF at the frontotemporal junction, insula and dmPFC (Lundström 2009, PET) | 5 | high | TV |
| 70 | de Groot et al., 2012 | double-blind experiment (sender–receiver): facial EMG, sniff measurement, eye-tracking, visual search | human | 10 male senders; 36 female receivers (Calvi 2020, Table 3) | Fear chemosignals induced fearful facial expressions and sensory acquisition: increased sniff magnitude, increased eye scanning; Disgust chemosignals induced disgusted facial expressions and sensory rejection: reduced sniff magnitude, reduced target-detection sensitivity, reduced eye scanning | 4 | some concerns | Abs |
| 71 | de Groot & Smeets, 2017 | meta-analysis (random-effects) + p-curve + p-uniform | human | 26 studies (2000–2017), N = 1,652 unique individuals (598 senders, 66 % male; 1,054 receivers, 69 % female); N… | Significantly right-skewed p-curve: half Z = −7.08, full Z = −7.31 (p < 0.0001); no sign of inadequate power (33 % power: Z = 3.60, p > 0.99); P-uniform detected no publication bias (r_low p = 0.439; r_high p = 0.751); no author reported… | 2a | some concerns | TV |
| 72 | Calvi et al., 2020 | systematic review (PRISMA), no meta-analysis | multi | 45 studies (1/1970–4/2019) from 1,228 records; interspecies: Siniscalchi 31 dogs, D'Aniello 40 dogs, Lanata 7 horses | Transmission of negative emotions (fear/anxiety/stress) has the strongest evidence; happiness less and less robust; women are more sensitive receivers; chemosignals from opposite-sex senders appear more effective; no differences by ethnicity | 3 | some concerns | TV* |
| 73 | Mujica-Parodi et al., 2009 | fMRI (original + independent replication) + odour perception + threatening-face discrimination behaviour (morph) | human | 144 first-time skydiving donors (pooled 2 batches × 20 donors); cortisol validation n = 40; original fMRI 16 receivers… | Stress confirmed: cortisol 0.229 → 0.584 µg/dL (t = 7.15, p < 0.001); condition × time interaction F = 39.87, p < 0.001; original fMRI STRESS > EXERCISE: left amygdala t = 4.80, Z = 3.68, p(SVC) = 0.02, MNI (−16, −10, −18), superficial/cortical amygdala | 4 | some concerns | TV |
| 74 | de Groot et al., 2020 | within-subject (fear vs neutral clip) + autonomic physiology + multivariate PLS-DA + measurement of volatile molecules… | human | Exp 1a: 40 male donors, N = 36 analysed (3 intensity groups n = 12); Exp 1b (PID): pooled samples from 3 donors | Fear vs neutral: more axillary sweat, higher SCL/SCR and breathing rate (p < 0.001), higher heart rate (p = 0.004); PLS-DA classified 100 % (Q² = 0.76, R² = 0.79); Leave-one-out intensity classification 30/36 = 83.33 % (chance 33 %): low 91.67 %, medium 75.00 %, high… | 4 | some concerns | TV |
| 75 | Sobel et al., 1998 | fMRI (sniffing with/without odour, smelling without active sniffing; Fourier analysis at sniff frequency) | human | NR (not accessible) | Sniffing (with or without odour) activated temporal piriform and the medial–posterior orbitofrontal gyrus — attributed to somatosensory stimulation by airflow; Smelling (regardless of sniffing) mainly activated lateral–anterior OFC | 4 | some concerns | Abs |
| 76 | Zelano et al., 2016 | iEEG in epilepsy patients + behavioural experiments in healthy participants (nasal vs oral breathing; phase… | human | iEEG 7 patients (5 with piriform electrodes) + P8; emotion task 70 recruited, analysed 21 nasal/17 oral/24… | Slow piriform oscillations correlated with respiration in all 5 patients (r = 0.64–0.87); amygdala did not (R = 0.01–0.25); hippocampus only in P1 (r = 0.77); Inspiration-phase-locked delta power increased in piriform (5/5), amygdala (7/7), hippocampus (7/7); reduced with oral breathing; piriform nasal r = 0.64 vs… | 4 | some concerns | TV |
| 77 | Folschweiller & Sauer, 2021 | narrative review (mechanisms, cross-species) | multi | NA (review) | Tracheotomy, nasal occlusion, OSN ablation (methimazole) and OB removal/inactivation all abolish RR in cortex and hippocampus; RR is local rather than volume-conducted; during freezing in mice, the mPFC is entrained at ~4 Hz; phase-specific inhibition of the mPFC at 4 Hz disrupts freezing (Dejean 2016); mPFC–BLA synchronise at 4 Hz (Karalis… | 5 | high | TV |
| 78 | Bagur et al., 2021 | animal experiment — in vivo electrophysiology + olfactory bulbectomy (OBX) + OB optogenetics +… | mouse | Male C57Bl6 mice; GAD-Cre 22 animals (15 ChR2, 7 GFP); OB–dmPFC synchrony n = 15; 100 dmPFC neurons; (STATUS r4 records… | During freezing, breathing converges to ~4 Hz; the OB shows prominent 4 Hz; OB–dmPFC synchrony at 3–6 Hz (p = 6.1e−5, n = 15); Granger OB→dmPFC p = 0.0353; OB removal abolishes dmPFC 4 Hz (p = 5.1e−4); ~45 % of dmPFC neurons phase-locked to 4 Hz during freezing vs ~23 % during activity (χ² = 9.82, p = 0.0017) | 5 | some concerns | TV |
| 79 | Daniels & Vermetten, 2016 | narrative review + paradigm proposal | human | NA (review); figures cited: Hinton 2004 N = 100 refugees; Abramowitz & Lichtenberg 2010 N = 36 patients… | Hinton 2004: 45 % of 100 refugees had an odour-triggered panic attack in the past month; 58 % of those had intrusive flashbacks; Vermetten 2007 (PET): diesel in veterans with PTSD increased rCBF in the amygdala, insula, mPFC and ACC and decreased it in the lateral prefrontal cortex (BA 22, 47) | 5 | high | TV |
| 80 | Rasch et al., 2007 | within-subject double-blind experiments (odor vs vehicle) + sleep fMRI | human | Exp.1 (SWS) n=18; Exp.3 (REM) n=17; Exp.5 (fMRI) n=12 (according to Table 1 of Hu 2020); n for the other experiments NR | Odour during SWS improved location memory: ≈97% vs ≈85–86% in controls (Nature News 97% vs 86%; Alzforum 97% vs 85%); Hu 2020: Hedges' g = 0.69, 95% CI [0.19, 1.18], Z = 2.717, p = 0.007; No effect on procedural memory (finger tapping, SWS): g = 0.08, CI [−0.52,… | 4 | some concerns | Abs |
| 81 | Arzi et al., 2012 | partial-reinforcement trace conditioning (tone–odor) during sleep + polysomnography +… | human | NR | During the night, tones alone elicited deeper sniffs if previously paired with a pleasant odour and shallower sniffs if previously paired with an unpleasant odour → new associative learning during sleep; the next morning the differential sniff response persisted even though participants were unaware of having learned | 4 | some concerns | Abs |
| 82 | Shanahan et al., 2018 | within-subject TMR during nap + simultaneous EEG–fMRI + MVPA searchlight | human | 32 recruited (21 female, mean age 25.25), 14 excluded (11 lacked SWS, 3 woke/noticed the odour) → n=18 analysed; analysis by… | Objects of the cued category were forgotten less than those of the uncued category: Wilcoxon Z=1.70, p=0.04 one-tailed; 15/18 participants benefited (binomial p=0.003); Category reactivation in vmPFC predicted the memory benefit: peak MNI [6, 46, −12], t(16)=7.53, p_FWE=0.01; temporal correlation peak… | 4 | some concerns | TV |
| 83 | Hu et al., 2020 | meta-analysis (three-level random-effects) | human | 73 articles/abstracts/datasets → 91 experiments, 111 independent samples, 212 effect sizes, N = 2,004; mean n/experiment 22;… | Overall TMR effect: g = 0.29, 95% CI [0.21, 0.38], Z = 6.711, p < 0.001; I² = 71%; Publication bias (Egger Z=8.489, p<0.001): trim-and-fill g = 0.18 [0.06, 0.30]; 3-parameter selection model g = 0.13 [0.06, 0.21]; Vevea–Woods severe one-tailed 0.05 (n.s.) | 1a | some concerns | TV |
| 84 | Chu & Downes, 2002 | between-subject double-cuing experiments (2 experiments) | human | Exp. 1: 42 (3 groups: label, congruent odour, incongruent odour; ≈14/group); Exp. 2: 40 (4 groups: label, incongruent odour… | Exp. 1: significant group effect on 6/7 scales, e.g. painful χ²=19.32, p<0.001; pleasant χ²=17.52, p<0.001; anxious χ²=13.11, p=0.0014; vivid χ²=13.08, p=0.0014; unique χ²=4.95, p=0.0842 (n.s.); The congruent-odour group changed most: pleasant 0.47; personal… | 4 | some concerns | TV |
| 85 | Willander & Larsson, 2006 | between-subject Galton–Crovitz cueing (word vs picture vs odor) | human | 93 older adults, 3 cue groups (≈31/group); exact ages NR | Odour-evoked memories were older, concentrated in the first decade of life (<10 years); word/picture cues peaked at ages 11–20; odour memories were accompanied by a stronger feeling of “being brought back to the past” and had been thought about/rehearsed less | 3 | some concerns | Abs |
| 86 | Larsson et al., 2014 | narrative review (Perspective) | human | NA (review; Herz 2004 fMRI n=5; n for Arshamian 2013 not stated) | LOVER framework: odour-evoked autobiographical memories are Limbic, Old, Vivid, Emotional, Rare; Herz 2004 (n=5): perfume odour activated the amygdala and hippocampus more than pictures; Arshamian 2013: odours activated the amygdala, hippocampus, piriform, entorhinal cortex and OFC; verbal cues recruited prefrontal… | 5 | NA | TV |
| 87 | Yeshurun et al., 2009 | within-subject fMRI + behaviour, 2 days (day 1, day 8) | human | 16 healthy participants (8 female), aged 24–36 | Day 8: objects assigned the first association 50.5 ± 3.8% vs second 34.5 ± 2.8% (F(1,15)=6.4, p<0.023); no difference olfactory vs auditory (F=0.03, p=0.85); Valence-dependent: unpleasant first association 56.3 ± 4.7% vs 28.8 ± 3.7% (t(15)=3.53, p<0.003); pleasant first association… | 4 | some concerns | TV |
| 88 | Gottfried et al., 2004 | event-related fMRI, cross-modal recognition memory | human | 19 recruited (10 female, mean age 25), 4 excluded → n=15; visual–visual control n=15; respiration measured in 15 other participants | Right posterior piriform (MNI 30, 3, −12; Z=3.34) and right anterior hippocampus (33, −24, −15; Z=2.83) were more active on correct recognition of old objects even though no odour was present at test; interaction with the visual–visual dataset in right piriform (Z=3.17) → olfaction-specific; respiration… | 4 | some concerns | TV |
| 89 | Dahmani et al., 2018 | cross-sectional behaviour + structural MRI (Study 1) + lesion case-control (Study 2) | human | Study 1: 60 recruited, 57 analysed for behaviour (24 spatial/33 response), 55 for volume (23/32), 54 for cortical thickness (23/31);… | Odour identification ↔ % landmarks found r=0.26 [0.06, 0.45]; ↔ 4/8 VM errors r=−0.37 [−0.52, −0.21] (n=57); Effect driven by spatial learners: r=−0.57 [−0.77, −0.33] (n=24); response learners r=−0.22 [−0.49, 0.14] (n.s.) | 3 | some concerns | TV |
| 90 | Igarashi et al., 2014 | animal experiment: multi-site tetrode recording (LFP + single-unit) during odour–place… | rat | 5 rats LEC + dCA1; 4 MEC; 3 pCA1–LEC; 4 new odours; 20 LEC–dCA1 pairs; 63–75 dCA1 cells and 76–82 LEC cells/session | 20–40 Hz power increased during odour sampling: dCA1 F(2,78)=118, P<0.001; LEC F(2,18)=67.1, P<0.001; MEC did not increase; 20–40 Hz LEC–dCA1 coherence increased progressively T1→T5 and was absent on error trials; local power did not change with learning (dCA1 F(4,36)=0.72, P=0.58) | 5 | some concerns | TV* |
| 91 | Lee et al., 2021 | animal experiment: optogenetic inhibition + opto-tagged single-unit recording + fiber… | mouse | Fan-cell inhibition n=10 mice; Wfs1 n=9; recordings of 213 fan cells/7 mice; dopamine inhibition n=9; SCH23390 n=5;… | Inhibition of LEC fan cells impaired learning of new odour pairs (ANOVA p=2.3×10⁻⁴; n=10); successful sessions 90.9% → 25.0% (binomial p=4.1×10⁻¹⁰); A/B recall unaffected; Inhibition of Wfs1+ pyramidal cells had no effect (p=0.60; n=9) | 5 | low | TV |
| 92 | Sullivan et al., 2000 | animal experiment: odour–shock conditioning + 2-DG autoradiography | rat | Rat pups PN8–PN12; 8–12/group (a), 8 (b), 5–6 (c, d), 6–8 (e) | With odour–shock pairing, PN9 pups developed an odour preference, PN10 and PN11 pups avoided the odour: group × age ANOVA F(2,54)=9.29, P<0.001; the shock was equally aversive at PN9 and PN10 (vocalisation, movement) | 5 | some concerns | TV |
| 93 | Moriceau & Sullivan, 2006 | animal experiment: odour–shock conditioning ± mother, 2-DG, intra-amygdala muscimol,… | rat | Rat pups PN8, PN12–15, PN21–23; n/group not stated (inferred from degrees of freedom ≈8 pups/group, e.g. F(3,28)) | PN12–15: mother absent → avoidance, mother present → preference with the same shock procedure: F(3,28)=25.563, P<0.0001; PN8 preference regardless of the mother (F(3,16)=3.917, P<0.005 according to the notes); PN21–23 avoidance regardless of the mother (F(3,15)=9.404, P<0.005) | 5 | some concerns | TV |
| 94 | Schaal et al., 2000 | non-randomized exposure-group study (newborn behaviour) | human | 24 full-term newborns (group whose mothers ate anise vs did not; number/group NR) | Infants in the anise group showed a stable preference for anise odour immediately after birth and on day 4; unexposed infants showed aversion or neutrality (abstract); fewer negative facial expressions in the exposed group: d = −1.72 (95% CI −2.906 to −0.527; p ≤ 0.005) (recalculated by PLoS ONE 2023) | 3 | high | Abs |
| 95 | Murphy, 2019 | narrative review | human | NA | Olfactory regions (olfactory bulb, anterior olfactory nucleus, transentorhinal/entorhinal Braak I–II, piriform, hippocampus) are where AD pathology appears earliest; odour memory and odour identification are affected in AD and in people at risk (APOE ε4) | 5 | NA | Abs |
| 96 | Devanand et al., 2015 | prospective community cohort (WHICAP), 2- and 4-year follow-up | human | 1,037 without dementia at baseline; 757 with ≥1 follow-up (498 without MCI, 129 non-amnestic MCI, 130 amnestic MCI… | Cognitive decline: RR 1.067/UPSIT point (1.040–1.095; p<0.0001); after covariates 1.065 (1.034–1.095); In people without MCI: UPSIT was predictive (RR 1.094; 1.057–1.133) whereas SRT-TR was not (RR 1.010; 0.983–1.037; p=0.478) | 2 | some concerns | TV |
| 97 | Olofsson et al., 2021 | narrative review | human | NA | Odour identification impairment rises from ~11–24% in middle age to 37–70% at age 70; many older people are unaware of it; odour memory is a separate factor distinct from verbal memory (2,280 people aged 60–100, Larsson 2016) | 5 | NA | TV |
| 98 | Herz, 2009 | critical review with selection criteria (author calls it 'systematic analysis';… | human | 18 studies (1998–2006) analysed in detail; individual-study n available only for Campenni 2004 n=90 female students… | Only 18 studies passed the selection criteria (excluded if violating ≥3/5 criteria; studies >10 years old, animal studies and clinical populations excluded); Campenni 2004 (n=90 female students): the “relaxing”/“stimulating” suggestion determined the direction of change in heart rate and skin conductance regardless of whether… | 5 | some concerns | TV* |
| 99 | Herz, 2016 | narrative review | human | NR (narrative review; Table 1 is an unreadable image); underlying study: Pinto 2014 >3,000 Americans aged ≥57 | People aged ≥57 with impaired odour identification had a ~4-fold higher risk of 5-year mortality (Pinto 2014, >3,000 people); odour-evoked memories cluster at ages 0–10 and are more emotional; the amygdala is more active when recalling odour-evoked memories (Herz 2004, fMRI); Arshamian 2013:… | 5 | some concerns | TV* |
| 100 | Ball et al., 2020 | systematic review (Cochrane) of RCTs, narrative synthesis without meta-analysis, GRADE | human | 13 RCTs, n=708 (sample sizes 18–186, median 66); agitation 10 RCTs/593; NPI 8 RCTs/346; AE 4/206; QoL 2/134; mood… | Agitation (10 RCTs, n=593; GRADE very low): of the 5 trials of moderate/low certainty, 4 found no effect, only Ballard 2002 showed benefit (CMAI MD −11.10; 95% CI −20.00 to −2.20); BPSD/NPI (8 RCTs, n=346): 4/5 trials of moderate/low certainty reported benefit… | 1a | high | TV* |
| 101 | Donelli et al., 2019 | systematic review + meta-analysis (7 MA, random-effects; PRISMA; PROSPERO CRD42019130126) | human | 90 studies in the qualitative synthesis (65 RCTs, 7,993 people; 25 NRS, 1,200 people); 37 RCTs in the quantitative synthesis (3,964 people) | Lavender inhalation, all scales, all controls (24 RCTs, n=1,682): Hedges g −0.73 [−1.00, −0.46], p<0.00001, I² 85%; only RCTs not at high RoB (n=535): g −0.64 [−0.82, −0.47], I² 0%; Inhalation was effective in high-anxiety situations (g −0.83) but NOT significant in situations… | 1a | some concerns | TV |
| 102 | Kasper et al., 2014 | multicentre, double-blind, double-dummy, 4-arm RCT, adaptive two-stage design (stopped at… | human | 539 randomised, 536 treated (Silexan 160 mg: 128; 80 mg: 135; paroxetine: 137; placebo: 136); FAS… | HAMA change (FAS, LOCF): −14.1 ± 9.3 (160 mg), −12.8 ± 8.7 (80 mg), −11.3 ± 8.0 (paroxetine), −9.5 ± 9.0 (placebo); Silexan 160 mg vs placebo: difference 4.0 (95% CI 1.9–6.1), p<0.001, Cohen d 0.50; 80 mg: 2.8 (0.9–4.8), p=0.003, d 0.37; paroxetine not… | 1b | some concerns | TV |
| 103 | Sánchez-Vidaña et al., 2017 | systematic review (no meta-analysis), Jadad quality | human | 12 RCTs, 1,226 people (984 female 80.3%; 224 male 18.3%; 18 unknown); cancer 682, pregnant women 333, menopause 90, depression… | 7/12 RCTs reported improvement in depressive symptoms; inhalation 2/5 positive, massage 5/8 positive; the highest-quality studies (Soden 2004 Jadad 5; Graham 2003, Sehhatie 2015 Jadad 4) were all negative; positive studies were mainly Jadad 1–2 | 1a | high | TV* |
| 104 | Lin et al., 2019 | systematic review + meta-analysis of RCTs (random-effects) | human | 31 RCTs; total n NR (not in the abstract) | Pooled effect of 31 RCTs: Hedges' g = 1.103 (95% CI 0.813–1.393), p<0.001; very high heterogeneity: Q = 341.544, p<0.001, I² = 91.216%; heterogeneity remained high on leave-one-out | 1a | high | Abs |
| 105 | Kiecolt-Glaser et al., 2008 | randomized controlled within-subject crossover (3 odors) × between-subject expectancy… | human | 56 healthy participants (21 male, 35 female), mean age 24.41 (SD 6.05; 18–43); 3 six-hour sessions each | Lemon increased PANAS positive affect more than lavender (adjusted p = 0.001) and water (p<0.004): time × odor F(6,240) = 4.43, p<0.001; consistent across 4 measures (PANAS, Stroop, LIWC, IAPS); lavender was no better than water on any mood measure; IAPS valence worse… | 1b | low | TV |
| 106 | Howard & Hughes, 2008 | double-blind placebo-controlled experiment, mixed design 3 (odor) × 3 (instruction) × 10… | human | 96 healthy female university students | Lavender odour had no effect on GSR during relaxation — no different from placebo or no odour; expectancy instructions had a paradoxical effect: told the odour would hinder relaxation → relaxed more (greater GSR decrease); told it would help → relaxed less | 1b | some concerns | Abs |
| 107 | de Araujo et al., 2005 | event-related fMRI + psychophysics + sniff control | human | 12 right-handed men, aged 23–35 (fMRI); 11 others for the sniff-control experiment | Test odour + “cheddar cheese” −0.10 ± 0.08 vs + “body odor” −0.86 ± 0.07; t(11) = 6.68, p ≪ 0.001; intensity unaffected by the label (0.55 vs 0.64); Clean air + “cheddar cheese” +0.02 ± 0.06 vs + “body odor” −0.40 ± 0.06; t(11) = 4.1, p<0.001 | 5 | some concerns | TV |
| 108 | Goel et al., 2005 | within-subject counterbalanced laboratory sleep study with polysomnography | human | 31 good sleepers (16 male, 15 female), aged 18–30 (mean 20.5 ± 2.4) | Lavender increased the proportion of slow-wave sleep (SWS) in both men and women; all participants reported higher morning “vigor” after the lavender night | 4 | some concerns | Abs |
| 109 | Perl et al., 2016 | polysomnography + air-dilution olfactometer, one-night laboratory study, between-subject… | human | 34 participants (35 recruited, 1 excluded; aged 22–36, mean 26.5 ± 2.5); NREM analysis 27/34 participants, 536/633 trials; total 1,010… | During NREM, delta increased by 9.79 ± 18.0%, t(26) = 3.22, P<0.005; slow spindles increased by 5.16 ± 7.81%, t(26) = 3.91, P<0.001; theta, fast spindles and beta unchanged; the increase exceeded that of odourless epochs (delta t(26) = 3.21, P<0.005) — not due to the natural trend of… | 4 | some concerns | TV |
| 110 | Hummel et al., 2009 | prospective controlled study, non-randomized, non-blinded | human | 56 patients with olfactory loss (40 training vs 16 not training); post-viral subgroup according to Yuan 2021: 24 vs 11 | The training group increased Sniffin' Sticks TDI scores relative to baseline; the non-training group was unchanged; the training group became more sensitive to the training odours (PEA threshold) | 2 | high | Abs |
| 111 | Sorokowska et al., 2017 | meta-analysis (random-effects, Hedges' g, meta-regression) | human | 13 studies (2009–2016); Identification k=13, n=1,005; Discrimination k=10, n=788; Threshold k=12, n=951;… | TDI: g = 1.10 (95% CI 0.459–1.734), p<0.001, k=11, n=879; Identification g = 0.83 (0.264–1.402); Discrimination g = 0.89 (0.498–1.298); Threshold g = 0.34 (0.103–0.569); post-infectious olfactory loss benefited most: TDI g = 1.52 vs 0.45 (Q = 4.34, p =… | 1a | some concerns | TV |
| 112 | Pieniak et al., 2022 | narrative review with summary table | multi | NR (narrative review; Table 1 summarises clinical OT trials) | Hummel 2009: 28% OT vs 6% no intervention after 12 weeks; Haehner 2013 (Parkinson’s): 20% vs 9%; Qiao 2019: improvement rate 1.7% after 1 month → 26.7% after 3 months → 41.7% after 6 months; Konstantinidis 2016: 58% (16 weeks), 71% (56 weeks), 37% no intervention | 5 | some concerns | TV* |
| 113 | Kattar et al., 2021 | systematic review + meta-analysis (PRISMA) | human | 1,981 abstracts → 16 articles in the review; meta-analysis of 4 controlled studies (total n NR) | OT vs control: OR = 2.77 (95% CI 1.67–4.58) for achieving a clinically meaningful TDI improvement (meta-analysis of 4 studies); 15/16 (93%) studies reported Sniffin' Sticks with full T, D, I and TDI; all studies reported clinically meaningful improvement after OT | 1a | some concerns | Abs |
| 114 | Kohli et al., 2016 | systematic review with pooled weighted means and weighted correlation (two populations) | human | Primary depression: 13 studies (10 poolable; e.g. SS Identification 152 depressed/208 controls); olfactory disorders… | Depression vs controls (pooled): SS threshold 6.31 vs 6.78 (p = 0.0005); discrimination 12.05 vs 12.66 (p = 0.0073); identification 12.57 vs 12.98 (p<0.0001); SIT-40 35.31 vs 37.41 (p<0.0001); BDI by olfactory status: normosmic 5.21 ± 4.73, hyposmic 10.93 ± 9.25, anosmic… | 3 | some concerns | TV |
| 115 | Athanassi et al., 2021 | narrative review (human + rodent models) | multi | NR (narrative review, no tables) | Hedonics is the most sensitive measure: unpleasant odours are rated more unpleasant and pleasant odours less pleasant (negative alliesthesia, Atanasova 2010); patients have difficulty detecting the pleasant component of odour mixtures (Naudin 2012); detection thresholds are inconsistent across studies… | 5 | some concerns | TV |
| 116 | Spangenberg et al., 1996 | pretest survey + experiment in simulated store (control + 2 affect × 3 intensity) | human | Pretest 704 people (26 essential oils, 23–31 people/odour); experiment 308 students recruited, 10 excluded → n = 298 (46%… | Scent improved store evaluations F(6,291) = 5.51, p<0.0001; environment F(14,283) = 3.50, p<0.0001; merchandise F(4,293) = 4.72, p<0.001; products F(18,279) = 2.34, p<0.002; intention to visit the store 4.20 → 4.94 (p = 0.003); intention to buy a backpack 2.29 → 2.97… | 4 | some concerns | TV* |
| 117 | Roschk & Hosseinpour, 2020 | meta-analysis (two-level HLM, Hunter–Schmidt correction, meta-analytic SEM/path analysis) | human | 71 independent samples from 64 articles (1989 – 5/2018), 15,447 people, 671 effect sizes; mean age 32.6; 60% female; 91% Western | Significant effects rₐ 0.078–0.177 (+3.2% to +15.4%); largest for recall rₐ = 0.177 (+11.3%); Spending: rₐ = 0.051, not significant (+3.2%; k=52, n=8,150) | 1a | low | TV |
| 118 | Krishna et al., 2010 | two laboratory experiments with pretests (between-subject; mediation analysis) | human | Exp. 1: 151 students (pretest 31); Exp. 2: 153 recruited → n = 143 (pretest 21) | Exp. 1: scent effect F(2,141) = 39.65, p<0.01, η² = 0.36; scent × delay interaction F(4,141) = 32.99, p<0.01; Recall (0 / 24 h / 2 weeks): no scent 3.27 / 1.94 / 0.87; familiar scent 4.71 / 4.83 / 3.62; unfamiliar scent 3.56 / 3.83 / 3.27 | 4 | some concerns | TV |
| 119 | Weber & Heuberger, 2008 | 4 field experiments (fragrant plants vs control outdoor setting) + 1 artificial odor… | human | Field Exp. 1–4: 32 healthy participants in total; Exp. 5: 22 participants | Plant scents were rated more intense than the control in all 4 field experiments; more pleasant in only 2/4; plant scents improved calmness, alertness and mood but depended on the order of conditions; independent of visual features | 4 | high | Abs |
TV = full text; TV* = partial full text; Abs = abstract only. RoB: summary assessment (see 2.5).
Scientific reference material, not medical advice. Figures are redrawn from published data, not copied from the original figures. GRADE and risk-of-bias ratings are the compilers’ own. Every number traces back to detailed reading notes on each paper.